Re: Again: (self)hosted Plan9? Was: [9fans] extending xen to allow
- writing drivers sucks. it's not a big problem in itself. i quite enjoy it for the reasonably well-documented chipsets one finds in (say) embedded ARM and PowerPC platforms. for those, i hardly ever bother to look at another driver. it's just so straightforward. i look at the book and do what it says. it doesn't work, so i find there's an errrata or fuss about discovering that a bit has the opposite sense from what's documented. no matter. on the PC, it's rather more troublesome: when i could get reasonable documentation it was much the same as anything else. without it, it's tedious, and perhaps too time-consuming if i'm doing it in my spare time. theo de raadt's slides were quite a good summary. still, there's not much choice, really.
Re: [9fans] IWP9 talks?
It takes a day for me to fly to either NY or Madrid. Get over it. Long haul sucks, but someone's gotta do it ... And Sydney in summer is more fun than NYC in winter. I did my usual panic get to work (Murray Hill) in a blizzard an didn't know about snow-day don't bother. There were 2 people that day in 1127, me and ken. brucee On 12/12/06, Paul Lalonde [EMAIL PROTECTED] wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Victoria, BC might be possible; but that's a lot of timezones for our european friends. Paul On 11-Dec-06, at 7:22 PM, ron minnich wrote: is canada easier for people? toronto or even somewhere really cool like newfoundland? ron -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.5 (Darwin) iD8DBQFFfjckpJeHo/Fbu1wRAltvAKCNMbtrBnsyLvXrmJ8lEYJAonZ5vwCeKf2o z2aTT9UFZ0UbTIXJQHXs3mw= =74MX -END PGP SIGNATURE-
Re: Again: (self)hosted Plan9? Was: [9fans] extending xen to allow
Browser, buy a cheap PC and run whatever you like on it. Wow that solves everything. That was a good session. brucee On 12/12/06, Charles Forsyth [EMAIL PROTECTED] wrote: - writing drivers sucks. it's not a big problem in itself. i quite enjoy it for the reasonably well-documented chipsets one finds in (say) embedded ARM and PowerPC platforms. for those, i hardly ever bother to look at another driver. it's just so straightforward. i look at the book and do what it says. it doesn't work, so i find there's an errrata or fuss about discovering that a bit has the opposite sense from what's documented. no matter. on the PC, it's rather more troublesome: when i could get reasonable documentation it was much the same as anything else. without it, it's tedious, and perhaps too time-consuming if i'm doing it in my spare time. theo de raadt's slides were quite a good summary. still, there's not much choice, really.
Re: Again: (self)hosted Plan9? Was: [9fans] extending xen to allow
Hello what books you guys recommend to start with hardware programming? (nemo's kernel book of course) I mean, having no experience with hardware programming, a desire i have is to read something to learn from other's experience on writing software for manage hardware. (something like the practice of programming but focused on hardware issues). of course i can always re-read my school notes, and start to fight with the real life. . . but this looks discouraging, (and becomes much more discouraging taking in account the comments of more talented programmers on the iwp9 :) thanks gabi On 12/12/06, Charles Forsyth [EMAIL PROTECTED] wrote: - writing drivers sucks. it's not a big problem in itself. i quite enjoy it for the reasonably well-documented chipsets one finds in (say) embedded ARM and PowerPC platforms. for those, i hardly ever bother to look at another driver. it's just so straightforward. i look at the book and do what it says. it doesn't work, so i find there's an errrata or fuss about discovering that a bit has the opposite sense from what's documented. no matter. on the PC, it's rather more troublesome: when i could get reasonable documentation it was much the same as anything else. without it, it's tedious, and perhaps too time-consuming if i'm doing it in my spare time. theo de raadt's slides were quite a good summary. still, there's not much choice, really.
Re: Again: (self)hosted Plan9? Was: [9fans] extending xen to allow
well while i'm commenting randomly ... none. there is essentially no documention on many, many boards and chips just in case we want to rip off their IP'. correct me if i'm wrong - i have stood corrected. brucee On 12/12/06, Gabriel Diaz [EMAIL PROTECTED] wrote: Hello what books you guys recommend to start with hardware programming? (nemo's kernel book of course) I mean, having no experience with hardware programming, a desire i have is to read something to learn from other's experience on writing software for manage hardware. (something like the practice of programming but focused on hardware issues). of course i can always re-read my school notes, and start to fight with the real life. . . but this looks discouraging, (and becomes much more discouraging taking in account the comments of more talented programmers on the iwp9 :) thanks gabi On 12/12/06, Charles Forsyth [EMAIL PROTECTED] wrote: - writing drivers sucks. it's not a big problem in itself. i quite enjoy it for the reasonably well-documented chipsets one finds in (say) embedded ARM and PowerPC platforms. for those, i hardly ever bother to look at another driver. it's just so straightforward. i look at the book and do what it says. it doesn't work, so i find there's an errrata or fuss about discovering that a bit has the opposite sense from what's documented. no matter. on the PC, it's rather more troublesome: when i could get reasonable documentation it was much the same as anything else. without it, it's tedious, and perhaps too time-consuming if i'm doing it in my spare time. theo de raadt's slides were quite a good summary. still, there's not much choice, really.
Re: Again: (self)hosted Plan9? Was: [9fans] extending xen to allow
there is essentially no documention on many, many boards and chips just in case we want to rip off their IP'. You're probably right. The only explanation is that these suckers actually know that their product can easily be improved upon. ++L
Re: Again: (self)hosted Plan9? Was: [9fans] extending xen to allow
I mean, having no experience with hardware programming, a desire i have is to read something to learn from other's experience on writing software for manage hardware. (something like the practice of programming but focused on hardware issues). possibly a good way is to read existing Plan 9 drivers it isn't really a deep mystery, except for some of the peculiar interfaces on the x86. i usually blunder my way past them, myself.
Re: [9fans] IWP9 talks?
And Sydney in summer is more fun than NYC in winter. yes, and surely 9fans isn't losing a sense of adventure...
Re: [9fans] IWP9 talks?
I was just wondering, because I'll be in NY in about a month and Murray Hill is on my agenda of places to visit - how closed are Labs nowadays? Just taking a walk outside is fine by me, but maybe there is a slight chance of getting inside under someone's supervision? In case there's still anything left over of 1127... Any idea? Boris On 11 Dec 2006, at 23:05, Joel Salomon wrote: On 12/11/06, Latchesar Ionkov [EMAIL PROTECTED] wrote: I'd vote for New York, but Austin is fine too :) Here's another vote for New York. (Especially if it can include a field trip across the Hudson to the land of New Jersey, where once dwelt the fair maidens UNIX and Plan 9. (Is anyone at Bell Labs still using Plan 9?)) Hope 7 sounds like a reasonably good time to meet in NY — especially since all other regional computer conferences seem to have moved to Boston — but not if it means waiting until 2009. --Joel
Re: Again: (self)hosted Plan9? Was: [9fans] extending xen to allow
- one significant place where Plan 9 wins is using it as a versatile base for building pieces that people use without knowing it's Plan 9 (e.g., Sape's wireless base stations, Rangboom, xcpu, and many Inferno apps that Charles can't talk about). I think we all want a bit more glory than that, but if we all convince ourselves, we can tuck Plan 9 entirely out of sight. Problem is, we then don't get a community any more and Plan 9 does not have a profit-making organisation that can support it without the community. Or am I missing something? It doesn't always have to be completely out of sight. Good embedded applications are free from the usual, `but everyone else is using windows' arguments. Just a minor point.
Re: Again: (self)hosted Plan9? Was: [9fans] extending xen to allow
Browser, buy a cheap PC and run whatever you like on it. It was few years ago when I realized that machines were cheap, take two. My secondary is a Mac.
Re: Again: (self)hosted Plan9? Was: [9fans] extending xen to allow
there is essentially no documention on many, many boards and chips just in case we want to rip off their IP'. You're probably right. The only explanation is that these suckers actually know that their product can easily be improved upon. ++L Another thing I've ran into was the small companies that thought it was easier to write drivers themselves than to write good documentation on their parts. The driver author just has to shout over the cube wall. They have to write the drivers anyway. Intel seems to be schizophrenic about it. Some docs we can get with a NDA and some are on the web. Not clear why. A recent driver was done with no docs but good support from the company. Some companies, LIKE BROADCOM!, won't sell part to Coraid because mocking high pitched voice we're too small /mocking high pitched voice. No docs, no support, hope they fall into a large hole.
Re: Again: (self)hosted Plan9? Was: [9fans] extending xen to allow
hello Some companies, LIKE BROADCOM!, won't sell part to Coraid because mocking high pitched voice we're too small /mocking high pitched voice. No docs, no support, hope they fall into a large hole. amen :) seems that the conclusion is what i 'suspected', reading the source and actually writting drivers is the only way to start thanks gabi
Re: Again: (self)hosted Plan9? Was: [9fans] extending xen to allow
On 12/12/06, Bruce Ellis [EMAIL PROTECTED] wrote: Browser, buy a cheap PC and run whatever you like on it. Wow that solves everything. That was a good session. brucee sucks for laptops though. I hate carrying all them thar laptops -- they get in the way of my shootin' iron. ron
Re: [9fans] IWP9 talks?
And Sydney in summer is more fun than NYC in winter. yes, and surely 9fans isn't losing a sense of adventure... Sydney would be good. for adventure, i'll put in a vote for Las Vegas. Vegas is setup for large groups and relatively inexpensive (if you don't gamble). Tempe, AZ (ASU campus) is a fun place too. You can wear your gun to the grocery store (can't be concealed); though a three-barreled pirate pistol may not be ok. In spring/summer you can't beat Seattle/Vancouver area. i plan to sponsor a plan9 gathering here soon.
[9fans] UBUNTU and v9fs
Does the wiki have instructions on how to make UBUNTU (Dapper Drake) speak v9fs? It seems to me that I need a recompiled kernel and I'm not familiar enough with the procedure to upgrade the system at that level. ++L
Re: Again: (self)hosted Plan9? Was: [9fans] extending xen to allow
On 12/12/06, ron minnich [EMAIL PROTECTED] wrote: On 12/12/06, Bruce Ellis [EMAIL PROTECTED] wrote: Browser, buy a cheap PC and run whatever you like on it. Wow that solves everything. That was a good session. brucee sucks for laptops though. I hate carrying all them thar laptops -- they get in the way of my shootin' iron. Just need to put your experience of building small systems towards building a headless laptop-server -- then you can use drawterm from your browser laptop - or home desktop, or whatever. It'd be sweet to have something I could power off of USB 2.0 or battery, with a hard drive and wireless (and maybe a serial port for jmk). Gumstick seems like it comes close - but no real solution for portable or piggy-back power. I suppose an iPaq might be able to be tasked to such a solution as well -- but has less than ideal disk storage. Neither has particularly glorious CPU power or memory -- maybe we can build something with the OLPC mother boards sans screen/keyboard. -eric
Re: [9fans] some snapshots of iwp9
ewww. How do we do anonymous 9p? What's wrong with sources? Too little disk capacity? ++L
Re: [9fans] some snapshots of iwp9
On 12/12/06, ron minnich [EMAIL PROTECTED] wrote: On 12/11/06, Lyndon Nerenberg [EMAIL PROTECTED] wrote: Just put 'em on an FTP server somewhere. Nobody on this list owns a web ^^^ ewww. How do we do anonymous 9p? contrib/papers on sources? I was thinking next workshop registration and paper submission should be a 9p-only affair. -eric
Re: [9fans] UBUNTU and v9fs
On Tue, Dec 12, 2006 at 05:35:58PM +0200, Lucio De Re wrote: Does the wiki have instructions on how to make UBUNTU (Dapper Drake) speak v9fs? It seems to me that I need a recompiled kernel and I'm not familiar enough with the procedure to upgrade the system at that level. You don't need to do anything, just ; modprobe 9p2000 or ; modprobe 9p should work fine... Regards, -- Anselm R. Garbe http://suckless.org/~arg/ GPG key: 0D73F361
Re: Again: (self)hosted Plan9? Was: [9fans] extending xen to allow
does the gumstix come with enough hardware documentation? perhaps i missed it. but i didn't see the info on how one boots these things. - erik On Tue Dec 12 10:20:29 EST 2006, [EMAIL PROTECTED] wrote: It'd be sweet to have something I could power off of USB 2.0 or battery, with a hard drive and wireless (and maybe a serial port for jmk). Gumstick seems like it comes close - but no real solution for portable or piggy-back power. I suppose an iPaq might be able to be tasked to such a solution as well -- but has less than ideal disk storage. Neither has particularly glorious CPU power or memory -- maybe we can build something with the OLPC mother boards sans screen/keyboard. -eric
Re: [9fans] UBUNTU and v9fs
On 12/12/06, Anselm R. Garbe [EMAIL PROTECTED] wrote: On Tue, Dec 12, 2006 at 05:35:58PM +0200, Lucio De Re wrote: Does the wiki have instructions on how to make UBUNTU (Dapper Drake) speak v9fs? It seems to me that I need a recompiled kernel and I'm not familiar enough with the procedure to upgrade the system at that level. You don't need to do anything, just ; modprobe 9p2000 or ; modprobe 9p Default Dapper has an older, buggier version of v9fs (2.6.15). Edgy has a much more reasonable version (2.6.17). -eric
Re: [9fans] some snapshots of iwp9
Still need to get some slides. When we got everything, we´ll put everything in the web site and also copy it into sources if that´s ok. (Most papers are already linked to the web page). On 12/12/06, Eric Van Hensbergen [EMAIL PROTECTED] wrote: On 12/12/06, ron minnich [EMAIL PROTECTED] wrote: On 12/11/06, Lyndon Nerenberg [EMAIL PROTECTED] wrote: Just put 'em on an FTP server somewhere. Nobody on this list owns a web ^^^ ewww. How do we do anonymous 9p? contrib/papers on sources? I was thinking next workshop registration and paper submission should be a 9p-only affair. -eric
Re: [9fans] IWP9 talks?
In spring/summer you can't beat Seattle/Vancouver area. Agreed. I was going to try and set something up here in Seattle if those other proposals don't work out.
Re: [9fans] UBUNTU and v9fs
Default Dapper has an older, buggier version of v9fs (2.6.15). Edgy has a much more reasonable version (2.6.17). Edgy+v9fs has been reliable. i have had a r/b client (uses v9fs) running continuously for a couple of weeks now.
Re: [9fans] UBUNTU and v9fs
On Tue, Dec 12, 2006 at 05:35:58PM +0200, Lucio De Re wrote: Does the wiki have instructions on how to make UBUNTU (Dapper Drake) speak v9fs? It seems to me that I need a recompiled kernel and I'm not familiar enough with the procedure to upgrade the system at that level. You don't need to do anything, just ; modprobe 9p2000 or ; modprobe 9p should work fine... Wow, does Ubuntu come with a 9p kernel module by default? That's sweet... /c
[9fans] system wide profile for rc(1)
Hello! while setting up my experimental Plan9 network I found myself drawn to implement a system wide 'profile' for the rc(1) shell. This allows to simplify both: rcmain and the personal 'lib/profile' for each user and seems easier to maintain to me. And .. it is so trivial that I must apologize for annoying the list with it, however maybe it is worth adopting it for the default installation. To make this mail small find it explained with details and ready to use sources at: http://www.magma.com.ni/moin/Plan9Tutorial/RcProfile Regards, Jorge-León
[9fans] (no subject)
Someone mentioned problems with getting upasfs faces to work with drawterm (I forgot who) I have put an annotated version of my profile in /n/sources/contrib/steve/rc/profile which does this fine. I use the same profile for diskless and diskfull terminals and drawterm. This uses the (IMHO) under-announced, and wonderfull feature of dt2k where it reads the users secstore into /dev/secstore. To do this use the -s command line option to drawterm to specify the secstore host you want to use. It is fairly complicated but it gets me an enviroment I like and it works everywhere. Hope its of some use. -Steve
Re: Again: (self)hosted Plan9? Was: [9fans] extending xen to allow
On 12/12/06, Eric Van Hensbergen [EMAIL PROTECTED] wrote: It'd be sweet to have something I could power off of USB 2.0 or battery, with a hard drive and wireless (and maybe a serial port for jmk). yeah, this was one of the ideas that came up before we started on Xen again, but Aki and Andrey and Lucho beat me up on this idea. It came down to the EC on one side, and me on the other, and that ended it. I suggested running linux on a little 1-5W board, and using it to run the linux apps, using a root mount from Plan 9. So Linux is this dumb little headless box you only turn on when you want, and otherwise you tell it to go away by yanking its power cord, verily. They thought the idea, uh, lacked merit. (I think they said it sucked, but am not sure). I think one reason the idea may really suck is that Firefox (the thin client) requires a 200 MB footprint, which translates to gobs of Watts. Figures. Web 2.0! [[BTW, anybody but me enjoying the idea of taking an opteron out of socket and replacing with ... an ... XML ... accelerator?]] But I still like the 'stupid little linux CPU' idea. I want a backpack full of little computers that spin up on demand. And don't weigh much. and take no power. And have no moving parts. And generate no heat. And use a fusion reactor for power. And, to reduce weight, have antigravity pods. I guess I'll go visit Area 52 this weekend (Area 51 is always behind schedule and over budget). ron
Re: Again: (self)hosted Plan9? Was: [9fans] extending xen to allow
This may solve the no-firefox-and-mplayer-for-plan9 problem, but I don't see how it solves the no-plan9-drivers-for-my-laptop one. Lucho On Dec 12, 2006, at 3:01 PM, ron minnich wrote: On 12/12/06, Eric Van Hensbergen [EMAIL PROTECTED] wrote: It'd be sweet to have something I could power off of USB 2.0 or battery, with a hard drive and wireless (and maybe a serial port for jmk). yeah, this was one of the ideas that came up before we started on Xen again, but Aki and Andrey and Lucho beat me up on this idea. It came down to the EC on one side, and me on the other, and that ended it. I suggested running linux on a little 1-5W board, and using it to run the linux apps, using a root mount from Plan 9. So Linux is this dumb little headless box you only turn on when you want, and otherwise you tell it to go away by yanking its power cord, verily. They thought the idea, uh, lacked merit. (I think they said it sucked, but am not sure). I think one reason the idea may really suck is that Firefox (the thin client) requires a 200 MB footprint, which translates to gobs of Watts. Figures. Web 2.0! [[BTW, anybody but me enjoying the idea of taking an opteron out of socket and replacing with ... an ... XML ... accelerator?]] But I still like the 'stupid little linux CPU' idea. I want a backpack full of little computers that spin up on demand. And don't weigh much. and take no power. And have no moving parts. And generate no heat. And use a fusion reactor for power. And, to reduce weight, have antigravity pods. I guess I'll go visit Area 52 this weekend (Area 51 is always behind schedule and over budget). ron
[9fans] abaco mailing list
Hola, Today starts my quest for a css and js compatible abaco, there are too many unanswered questions and design issues that need to be adressed , if you are interested you know what to do. to subscribe send an email to [EMAIL PROTECTED] or go to http://groups.google.com/group/websucks thanks -- Federico G. Benavento
Re: [9fans] system wide profile for rc(1)
Hello, I'm not quite sure what you want to achieve there (well, maybe I'm just slow). The net gain I see is two extra lines in rcmain and one additional file (actually two files) doing more or less the same as before. The two main usage scenarios I can think of would be a system used mainly by a single person and a system (or network of systems) used by a potentially large and diverse group of users. For a singleuser system that separation is more or less useless. For a multiuser system it might look like a good idea at first, but I value the convenience of having all settings I might want to customise in one place (especially as I get it via newuser anyway). The newuser script arguably is simpler, but the change really hasn't got anything to do with a system-wide profile and as a downside makes it no longer self-contained. All in all it's a little exercise in establishing a system policy for those who want it. Speaking for myself, it would certainly make me sad to see the rise of the POSIX-rc... Martin
Re: Again: (self)hosted Plan9? Was: [9fans] extending xen to allow
On 12/12/06, Latchesar Ionkov [EMAIL PROTECTED] wrote: This may solve the no-firefox-and-mplayer-for-plan9 problem, but I don't see how it solves the no-plan9-drivers-for-my-laptop one. Hence the approach we are taking with xen. You get a linux, you get a Plan 9, you get holes torn in the I/O and memory spaces from Plan 9 to hardware to let you doink hardware and write drivers under Plan 9, and, with any luck, crash the machine at will. I think it's going to work out. At least the crashing part. Aki is already regularly crashing linux from user mode, so how hard can it be? ron p.s. with the new web 2.0 in my firefox browser, with 64M resident out of 136M, I am seeing that the mouse sticks and as I move the mouse random shit gets highlighted and erased. Fun.
Re: [9fans] system wide profile for rc(1)
One could also simply . /rc/lib/profile at the top of the user's lib/profile if conformance was desired. Russ
Re: Again: (self)hosted Plan9? Was: [9fans] extending xen to allow
On Tue, Dec 12, 2006 at 09:30:35PM +1100, Bruce Ellis wrote: i thought that i could get anything when working at the labs. oh no. Wasn't there one time when there was problem getting drivers for *lucent* wavelan pcmcia cards? (Maybe I misremember, but the story is better that way.)
Re: Again: (self)hosted Plan9? Was: [9fans] extending xen to allow
Hence the approach we are taking with xen. You get a linux, you get a Plan 9, you get holes torn in the I/O and memory spaces from Plan 9 to hardware to let you doink hardware and write drivers under Plan 9, and, with any luck, crash the machine at will. Have your considered inverting this setup? Rather than a native Linux and a parasitic plan9, have a native plan9 hand over io and memory space it doesn't understand to the parasite. I'd rather have a very lean, clean and thin native os (AKA hypervisor). Of course I have no idea if this can be made to work
Re: [9fans] system wide profile for rc(1)
* Russ Cox ([EMAIL PROTECTED]) wrote: One could also simply . /rc/lib/profile at the top of the user's lib/profile if conformance was desired. Russ I also thought about that after posting. Has the additional advantage of not forcing the view of the site admin on you. Martin
[9fans] Plan9 under parallels; kernel rebuild
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 I've been following David Leimbach's directions on getting plan9 to run under parallels, and have run into a glitch while rebuilding the kernel. At link I get: configure: undefined: pcibussize in configure Any quick clues as to what's up? It's a recent plan9 iso - about 2 weeks old. Paul -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.5 (Darwin) iD8DBQFFf0b3pJeHo/Fbu1wRAsAPAJ0UYwAfsIJqzk39qfbEi94pWt9gpgCggmK8 wM4r9Hh68/NAvxaT7V8UseE= =gRh0 -END PGP SIGNATURE-
Re: [9fans] system wide profile for rc(1)
Martin Neubauer [EMAIL PROTECTED] writes:
...
main usage scenarios I can think of would be a system used mainly by a
single person and a system (or network of systems) used by a potentially
large and diverse group of users. For a singleuser system that separation
is more or less useless.
Right!
For a multiuser system it might look like a good
idea at first, but I value the convenience of having all settings I might
want to customise in one place (especially as I get it via newuser anyway).
As a system administrator I (may) have three requirements:
- set up several different networks of systems, each of which has its
own particularities
- change some particularities for all users within a certain network
of systems.
- Take care of users that are not (shell) programmers
for the first two requirements it is convenient to be able to change default
values for all users:
Default installation remains identical for all networks of systems, to
brand each network, only one /rc/lib/profile has to be changed, no
need to change ~/lib/profile foreach ~.
What is left in /sys/lib/rcmain is a bare minimum for running rc
non-interactively (I guess even the switch($#prompt.. could be
omited.
Also ~/lib/profile should be a bare minimum. See PS.1 for more.
/sys/lib/profile gives a reasonable default working environment
available even with a broken ~/lib/profile.
(l)users do break things. I want:
1- a working namespace without ~/lib/profile
2- a minimal ~/lib/profile, so user is not easily confused
3- be able to overwrite a broken, or restore a deleted ~/lib/profile
from a known to work template
1- and 2- are covered by /rc/lib/profile
The newuser script arguably is simpler, but the change really hasn't got
anything to do with a system-wide profile and as a downside makes it no
longer self-contained.
3- is covered by not self-containing ~/lib/profile in the newuser
script, but rather have it separate.
The same would apply to ~/lib/plumbing, however I have not
exercised it.
All in all it's a little exercise in establishing a system policy for those
who want it. Speaking for myself, it would certainly make me sad to see the
rise of the POSIX-rc...
Martin
I don't care about POSIX-rc or not, I need a simple yet scalable
system. This means (amongst others), that the general case must
include the special case.
In the given approach /sys/lib/profile can simply be omitted and
everything works as before.
The exercise tries to draw from the idea of separating concerns:
- rcmain sets up rc(1)
... only with rc -l:
- system wide profile sets up the default namespace
- user profile launches the programs each users wants to launch
The system wide profile should not do anything, the user profile
cannot undo or override later.
Regards,
Jorge-León
PS.1:
~/lib/profile could be stripped down to something like:
upas/fs
switch($service) {
case terminal
plumber
exec rio -i riostart
case cpu
news
if (! test -e /mnt/term/mnt/wsys) { # cpu call from drawterm
exec rio
}
case con
news
}
if /rc/lib/profile was extended with:
fn cd { builtin cd $* awd } # for acme
switch($service){
case terminal
echo -n accelerated '#m/mousectl'
echo -n 'res 3' '#m/mousectl'
case cpu
if (test -e /mnt/term/mnt/wsys) { # rio already running
bind -a /mnt/term/mnt/wsys /dev
if(test -w /dev/label)
echo -n $sysname /dev/label
}
bind /mnt/term/dev/cons /dev/cons
bind /mnt/term/dev/consctl /dev/consctl
bind -a /mnt/term/dev /dev
if (! test -e /mnt/term/mnt/wsys) { # cpu call from drawterm
font=/lib/font/bit/pelm/latin1.8.font
}
case con
news
}
PS.2: To illustrate the advantage for system administration of large
Networks consider the following ficticious scenario:
I have one fileserver with comercial software and shared storage for
the bookkeepers folks (3 persons).
For developers group one I have some special development software on a
second fileserver (7 persons).
For developers group two I have other software and their filespace (4 persons).
For testers group one I have a test-setup with application binaries
from developers group one fileserver (4 persons).
etc...
So I would change 9fs, mounts and binds (to /bin) in the
/rc/lib/profile of each group: i.e. four (4) files.
With the actual approach I would need to temper with 3+4+4=11
~/lib/profile files to get things right for the folks *and* I'd need
to have access rights to these: This I do not want at all!
Some three months later, the testers get bumped up to 7 persons and
need to test software from developers group two also:
Actual situation:
- create three new accounts, edit 4 old and 3 new ~/lib/profiles: 7 files
- alternative: edit ~/sys/lib/newuser and create the new accounts
afterwards:
Re: [9fans] Plan9 under parallels; kernel rebuild
keep in mind that the source that runs well on parallels is not the same as the source on sources. I don't know if the BIOS changes JMK did ever got merged in. Dave On 12/12/06, Paul Lalonde [EMAIL PROTECTED] wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 I've been following David Leimbach's directions on getting plan9 to run under parallels, and have run into a glitch while rebuilding the kernel. At link I get: configure: undefined: pcibussize in configure Any quick clues as to what's up? It's a recent plan9 iso - about 2 weeks old. Paul -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.5 (Darwin) iD8DBQFFf0b3pJeHo/Fbu1wRAsAPAJ0UYwAfsIJqzk39qfbEi94pWt9gpgCggmK8 wM4r9Hh68/NAvxaT7V8UseE= =gRh0 -END PGP SIGNATURE-
quite Off Topic: Re: Again: (self)hosted Plan9? Was: [9fans] extending xen to allow
ron minnich [EMAIL PROTECTED] writes: On 12/12/06, Eric Van Hensbergen [EMAIL PROTECTED] wrote: It'd be sweet to have something I could power off of USB 2.0 or battery, with a hard drive and wireless (and maybe a serial port for jmk). yeah, this was one of the ideas that came up before we started on Xen again, but Aki and Andrey and Lucho beat me up on this idea. It came down to the EC on one side, and me on the other, and that ended it. Wikipedia: Early Childhood education Electric Circus Elimination Chamber Emergency Contraception Eric Clapton Exacoulomb ahh EC = Executive Committee! 8-0 I suggested running linux on a little 1-5W board, and using it to run the linux apps, using a root mount from Plan 9. So Linux is this dumb little headless box you only turn on when you want, and otherwise you tell it to go away by yanking its power cord, verily. They thought the idea, uh, lacked merit. (I think they said it sucked, but am not sure). I think one reason the idea may really suck is that Firefox (the thin client) requires a 200 MB footprint, which translates to gobs of Watts. Figures. Web 2.0! [[BTW, anybody but me enjoying the idea of taking an opteron out of socket and replacing with ... an ... XML ... accelerator?]] But I still like the 'stupid little linux CPU' idea. I want a backpack full of little computers that spin up on demand. And don't weigh much. and take no power. And have no moving parts. And generate no heat. And use a fusion reactor for power. And, to reduce weight, have antigravity pods. I guess I'll go visit Area 52 this weekend (Area 51 is always behind schedule and over budget). http://www.intellasys.net/ Though: Do you feel like re-implement Plan9 in machine Forth? ron -- Jorge-León
Re: [9fans] system wide profile for rc(1)
Maybe I'm missing something, but why not just modify /rc/bin/termrc or termrc.local?
Re: [9fans] Plan9 under parallels; kernel rebuild
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 I dug around a bit, and duplicated an old pcibussize function (that just passed through to a pcibusmap and life is good. I guess I could have re-written the (one) call point just as trivially. Runs now, and sees the outside world. Now to turn it into a CPU server. Paul On 12-Dec-06, at 4:33 PM, David Leimbach wrote: keep in mind that the source that runs well on parallels is not the same as the source on sources. I don't know if the BIOS changes JMK did ever got merged in. Dave On 12/12/06, Paul Lalonde [EMAIL PROTECTED] wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 I've been following David Leimbach's directions on getting plan9 to run under parallels, and have run into a glitch while rebuilding the kernel. At link I get: configure: undefined: pcibussize in configure Any quick clues as to what's up? It's a recent plan9 iso - about 2 weeks old. Paul -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.5 (Darwin) iD8DBQFFf0b3pJeHo/Fbu1wRAsAPAJ0UYwAfsIJqzk39qfbEi94pWt9gpgCggmK8 wM4r9Hh68/NAvxaT7V8UseE= =gRh0 -END PGP SIGNATURE- -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.5 (Darwin) iD8DBQFFf01QpJeHo/Fbu1wRAiXRAKCaqMtOCxGmiTLRAMaGnZGwPupHoACfTB7l wCGcs+robSnUApdS9riBqpg= =OIq9 -END PGP SIGNATURE-
Re: Re: [9fans] IWP9 talks?
On 12/12/06, Tim Wiess [EMAIL PROTECTED] wrote: In spring/summer you can't beat Seattle/Vancouver area. Agreed. I was going to try and set something up here in Seattle if those other proposals don't work out. Seattle I could actually attend as I live here. Sorry I've absolutely no sense of adventure. But then again, I'd probably only be able to participate as an interested outsider rather than a full blown contributer. Life's been exceptionally fun lately.
Re: Re: [9fans] Plan9 under parallels; kernel rebuild
Let me know how it turns out! :-) I'd be glad to know it's maintainable beyond first installation. I've honestly not tried it, and started using a much newer drawterm to connect to a CPU box I put together (AMD Athlon 2800+), it's pretty sweet. Dave On 12/12/06, Paul Lalonde [EMAIL PROTECTED] wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 I dug around a bit, and duplicated an old pcibussize function (that just passed through to a pcibusmap and life is good. I guess I could have re-written the (one) call point just as trivially. Runs now, and sees the outside world. Now to turn it into a CPU server. Paul On 12-Dec-06, at 4:33 PM, David Leimbach wrote: keep in mind that the source that runs well on parallels is not the same as the source on sources. I don't know if the BIOS changes JMK did ever got merged in. Dave On 12/12/06, Paul Lalonde [EMAIL PROTECTED] wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 I've been following David Leimbach's directions on getting plan9 to run under parallels, and have run into a glitch while rebuilding the kernel. At link I get: configure: undefined: pcibussize in configure Any quick clues as to what's up? It's a recent plan9 iso - about 2 weeks old. Paul -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.5 (Darwin) iD8DBQFFf0b3pJeHo/Fbu1wRAsAPAJ0UYwAfsIJqzk39qfbEi94pWt9gpgCggmK8 wM4r9Hh68/NAvxaT7V8UseE= =gRh0 -END PGP SIGNATURE- -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.5 (Darwin) iD8DBQFFf01QpJeHo/Fbu1wRAiXRAKCaqMtOCxGmiTLRAMaGnZGwPupHoACfTB7l wCGcs+robSnUApdS9riBqpg= =OIq9 -END PGP SIGNATURE-
Re: Re: Again: (self)hosted Plan9? Was: [9fans] extending xen to allow
after watching andrey run plan9 in parallels and seeing the drawterm windows on x11, i'd have agreed with you, but i must say it feels a lot less bad when the parasite is full-screen and the host programs like firefox are in a smaller rio window that you can kill or hide at will. the firefox i'm writing this in is running full-screen (f11) on a 1000 by 1000 pixel vnc session and the illusion of having it native is pretty damn good (for me anyway) at least until my drawterm hack crashes and i start shouting untranslatable words (i hope it's better now that i realized drawterm didn't have reentrant memimagedraw) On 12/12/06, David Leimbach [EMAIL PROTECTED] wrote: I can't deny the utility of having Firefox (I'm writing this in a Firefox window), but even if Plan 9 could run Firefox, the next thing would be oh but it needs to be able to run these ten plugins, and so on and so on. Personally, I think you are going to be much happier running Plan 9 in some VM environment on Linux or Windows than putting in the effort for the other way around. And a lot of times, at the end of the day, I feel that as a result of wanting to run Plan 9 in a VM environment, even in Parallels, makes me sad, and I'd almost rather use Inferno :-) Russ
Re: Re: [9fans] IWP9 talks?
Seattle I could actually attend as I live here. Sorry I've absolutely no sense of adventure. people in Seattle and vicinity, anyone interested in getting a 9fans monthly meeting together? downtown someplace perhaps?
Re: Re: [9fans] IWP9 talks?
I'm not in Seattle, but 2600 meetings are good things to meet people. And if Plan9-ers are in your area thats a meeting you could all attend. Monthly, first firday of every month, in most places of the world. http://2600.com /random On 12/12/06, Skip Tavakkolian [EMAIL PROTECTED] wrote: Seattle I could actually attend as I live here. Sorry I've absolutely no sense of adventure. people in Seattle and vicinity, anyone interested in getting a 9fans monthly meeting together? downtown someplace perhaps?
[9fans] Interrupted alarm(2)
I’ve been simulating preemptive multithreading using alarm(2) notes
and lots of stack smashing. (I’ll have some questions about that
later, if things don’t go smooth.)
Anyhow, my taskfork looks like:
void
taskfork(ulong flags)
{
longquantum;
...
quantum = alarm(0); // stop timer and save time left
...
// back to our regularly scheduled programming...
alarm(quantum == 0 ? 1 : quantum);
}
I included that ugly last line on the off chance that taskfork() is
called with less than a millisecond left and so alarm(0) will return
0. Is this neccessary, or can alarm(0) not ever return 0? Would it
be better to write:
if(quantum)
alarm(quantum);
else
postnote(PNPROC, getpid(), alarm);
}
--Joel
Re: quite Off Topic: Re: Again: (self)hosted Plan9? Was: [9fans] extending xen to allow
On 12/12/06, Georg Lehner [EMAIL PROTECTED] wrote: ron minnich [EMAIL PROTECTED] writes: It'd be sweet to have something I could power off of USB 2.0 or battery, with a hard drive and wireless (and maybe a serial port for jmk). But I still like the 'stupid little linux CPU' idea. Something along these lines? http://www.engadget.com/2005/08/11/blackdog-linux-a-em-real-em-pocket-pc/ (Sorry, haven't been following the thread) -Jack
Re: [9fans] Interrupted alarm(2)
Use the source, Luke! // back to our regularly scheduled programming... alarm(quantum == 0 ? 1 : quantum); } I included that ugly last line on the off chance that taskfork() is called with less than a millisecond left and so alarm(0) will return 0. Is this neccessary, or can alarm(0) not ever return 0? Would it be better to write: Certainly I would guess that alarm(0) returning 0 is a perfectly reasonable thing to do, if the alarm was either imminent or not scheduled in the first place. System calls are implemented in /sys/src/9/port/sys*.c, by functions named sysfoo for the call foo. In this case sysalarm just calls procalarm, which is in /sys/src/9/port/alarm.c. Inspecting that, it is clear that alarm(0) can in fact return 0 in the two cases I just mentioned. Perhaps more worrying, it looks like it can also return a negative number if the alarm time has passed without the alarm actually going off. And it might not even be the right negative number. I also don't see anything that keeps one from scheduling alarms at times in the past, though they will just be treated as happening as soon as possible (e.g., alarm(-100) looks like it will cause an alarm to go off at the next clock tick). It also looks like the delivery of alarms can be stopped by a single rogue process that can manage to make its p-debug qlock unlockable and then schedule a quick alarm to get to the front of the alarm queue. Making p-debug unlockable is easy: one way is to arrange that the demand load fault caused by memmove in case Qnote in procread causes i/o to a file server that sits on the request. Then there will be no more alarms delivered anywhere in the system until that i/o finishes. This may be less than ideal. Then again, the whole idea of alarms is broken if you are depending on them for very much at all, so the fact that the implementation is also a little broken probably doesn't matter. Russ
[9fans] IPW10.ca
On Dec 11, 2006, at 7:22 PM, ron minnich wrote: is canada easier for people? toronto or even somewhere really cool like newfoundland? Thetis Island. While I'm not (yet) on the excluded list, the concept of bending over for TSA just doesn't do it for me any more. Thetis is a four hour sail from Vancouver, has reasonably non-stale Guinness and single malt, and the girls there are so adventurous that even Plan 9 geeks might stick around and get married :-) --lyndon
