Linux-Development-Sys Digest #825, Volume #6 Sat, 12 Jun 99 15:14:12 EDT
Contents:
Re: Pinning a thread to a processor (Aaron Coday)
Re: Configuration Manager for Linux (Leslie Mikesell)
Re: egcs bootstrap (Paul Kimoto)
Re: Detecting multiple keystrokes simultaneously (Roland Latour)
Re: Configuration Manager for Linux (Jonathan Abbey)
Re: Questions on Linux Memory Managment ([EMAIL PROTECTED])
Have you used QNX ? (D Allinson)
ORB for Linux w/ Java Binding? ("Mark A. Richman")
Re: Testing a kernel, boot from floppy? (Medical Electronics Lab)
Re: Need help on /proc/[pid]/* (George MacDonald)
Re: TAO: the ultimate OS (Matthias Buelow)
smbios (DMI) support in Linux? (Neil Schemenauer)
Re: Linux & Cybercafe (David Knight)
Re: Casio QV-11 on Lava DSerial PCI Port, follow-up (Marco Fonseca)
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (Aaron Coday)
Subject: Re: Pinning a thread to a processor
Date: 11 Jun 1999 08:37:22 -0700
So as far as I know there is no way as Linux exists now to bind the
process to the processor. I worked on a project, where we made a
modification to the kernel to allow this. Unfort., I can't release
the source on that one. It's not that hard a modification if you're
comfortable modifying a kernel. You basically have to add a system
call to be able to specify the binding, and then a couple of state
variables to the struct_task. Then you can make the schedule handle
the real work in the can_schedule function.
-Aaron
"Karsten Scholtyssik" <[EMAIL PROTECTED]> writes:
>bill davidsen <[EMAIL PROTECTED]> schrieb in im Newsbeitrag:
>7jpdgk$d90$[EMAIL PROTECTED]
>> In article <7jmjk5$ih9$[EMAIL PROTECTED]>,
>> Karsten Scholtyssik <[EMAIL PROTECTED]> wrote:
>>
>> | I'm looking for a method to set the CPU affinity of a process/thread.
>Under
>> | Windows NT, I use the function SetProcessAffinityMask() - what is the
>Linux
>> | and/or Solaris equivalent? (I hope there *is* a way at all)
>> |
>> | I searched the man pages (pthread etc.) to no avail.
>>
>> Out of curiousity, what do you think this binding with do for you?
>It prevents the process from beeing scheduled on a different CPU
>which increases the cache locality. It reduces the situations where
>the process has to start with a cold cache.
>> Also, does this NT call bind the process to the processor or the
>> processor to the process?
>It binds the process to the processor so that the process is only
>scheduled on the given Processor(s).
> Karsten
------------------------------
From: [EMAIL PROTECTED] (Leslie Mikesell)
Subject: Re: Configuration Manager for Linux
Date: 11 Jun 1999 11:30:24 -0500
In article <7jq6qg$[EMAIL PROTECTED]>,
Jonathan Abbey <[EMAIL PROTECTED]> wrote:
>I don't make any claims as to the goodness or reliability of the
>Ganymede database as compared to PostgreSQL, certainly. I will say
>that PostgreSQL itself doesn't do everything that Ganymede does, and
>that anything sophisticated that you'd put together using PostgreSQL
>would have to go through the same development process that Ganymede
>has already gone through at this point, especially if you were
>committed to having all changes go through an intelligent management
>layer instead of having a more ad hoc repository of information that
>various tools collaborated on in a less sophisticated way.
PostgreSQL does have server-side programmability using triggers
but it would also be reasonable to use it only as a data store
with updates forced to go through the front end that knows about
the additional constraints that you would like to enforce.
You could still do things like point Crystal Reports at some
of the tables via ODBC when you need a nice pretty printout
with arbitrary grouping levels and detail lines.
>| [Wonderful sounding PostgreSQL features described ...]
I realize that this is easy to say with hindsight and don't mean
to be critical about decisions made when the feature set was much
more limited. However, it is somewhat predictable that open
source general-purpose software will develop faster than specialized
things with a smaller user base and I expect that trend to continue.
If you pushed the data store off onto some other well tested code
it might be easier to concentrate on the unique features.
>True as well. That's why Ganymede is being sold as a thing to feed
>NIS and DNS and LDAP and other things for which there is already
>well-defined API's and client code.
The problem I see with this is that most of these also define
network API's for updates (maybe all with the new dynamic DNS
stuff coming around). If you are feeding static data from
some other source you have to give up any native update protocols.
>With LDAP, there are more problems. LDAP's data types and the way one
>structures directory services with LDAP doesn't necessarily map into
>Ganymede's objects-and-links world very well.
LDAP doesn't even map to LDAP very well, given that there is no
transactional mode to create an object and it's parent in
an atomic operation. But, if the data can be stored in a
dbm file, anything else should be able to handle it.
>... I already have my hands quite full getting rich DNS
>support developed.
Same issues, if you want dynamic updates you have to provide
a front end that emulates the network API for the existing
clients. The same applies to all the existing password changing
mechanisms.
>As long as external code can be connected to the Ganymede server to
>support things like NT domain control and LDAP queries, I'm personally
>going to have to leave off supporting those things natively in favor
>of proxies and auxiliary servers fed by Ganymede.
Proxies would be great. I'd even be happy to allocate some IP addresses
to the proxies if the ports for the real services can't be moved.
Maybe it would be possible to pass queries through to the native
services but intercept updates and propagate to everything else
that needs them.
Les Mikesell
[EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Paul Kimoto)
Subject: Re: egcs bootstrap
Date: 11 Jun 1999 11:50:02 -0500
Reply-To: [EMAIL PROTECTED]
In article <7jr80k$am8$[EMAIL PROTECTED]>, gene wrote:
> I think I deleted something and now get this when I try to compile some
> programs:
>
> /tmp/ccPpUPmM.o: In function `main':
> /usr/src/gatos/yuvsum.c:64: undefined reference to `lstat'
> collect2: ld returned 1 exit status
> make: *** [yuvsum] Error 1
>
> and various 'stat' functions
As a stab in the dark, are you having the stat problem described in the
Glibc FAQ (http://sourceware.cygnus.com/glibc/glibc-faq.html)? (Please
disregard if you are running a libc5 system.)
> I wanted to make a bootstrap compile of egcs-1.1.2 and get the same
> errors. In the 'BUILD' document it says:
>
> For a native build issue the command "make bootstrap". This will build
> the entire egcs compiler system, which includes the following steps:
[...]
> I would like to try this but the egcs docs give no instruction about the
> proper way to do it.
The proper way to do what?
Again stabbing in the dark: you need to configure before running "make".
The installation instructions are (duplicated?) on the web at
http://egcs.cygnus.com/install/index.html .
--
Paul Kimoto <[EMAIL PROTECTED]>
------------------------------
From: Roland Latour <[EMAIL PROTECTED]>
Subject: Re: Detecting multiple keystrokes simultaneously
Date: Fri, 11 Jun 1999 08:49:32 -0700
Brian Luczkiewicz wrote:
>
> I am writing a program that needs to deal with the interpretation of
> multiple keystrokes at the same time such as Ctrl-Shift-A or similar. I
> think that to do this, you need to set up a timer or at least a loop, but
> wouldn't that be processor hungry? I am pretty new at this, but I think a
> timer would work. I don't think so, but would I need RTlinux for this?
This doesn't work the way you seem to think it does.
This is from 'man ascii':
Oct Dec Hex Char
101 65 41 A
141 97 61 a
115 77 4D M
015 13 0D CR
So 'a' = 0100 0001
'A' = 0110 0001
The caps key sets one bit high.
'm' = 0100 1101
CR = 0000 1101
CR is control-M, which you may have seen as pesky end-of-line on
some files from that other OS. The control key grounds one bit.
Of course, this is after the OS has processed keycodes into characters,
but that's what your program will see. I hope this helps.
--
Retired TechSupport Engr. Linux@CDSnet:http://home.cdsnet.net/~rolandl
"The urge to save humanity is almost always a false front for the
urge to rule." -H. L.Mencken
------------------------------
From: [EMAIL PROTECTED] (Jonathan Abbey)
Subject: Re: Configuration Manager for Linux
Date: 11 Jun 1999 12:03:28 -0500
In article <7jrdj0$2r7l$[EMAIL PROTECTED]>,
Leslie Mikesell <[EMAIL PROTECTED]> wrote:
|
| PostgreSQL does have server-side programmability using triggers
| but it would also be reasonable to use it only as a data store
| with updates forced to go through the front end that knows about
| the additional constraints that you would like to enforce.
| You could still do things like point Crystal Reports at some
| of the tables via ODBC when you need a nice pretty printout
| with arbitrary grouping levels and detail lines.
That does sound awfully nice, but getting from here to there would be
a significant research/development project unto itself (the problem of
schema editing/evolving would still have to be solved), and I don't
have the resources to do it. I wish I did.
| I realize that this is easy to say with hindsight and don't mean
| to be critical about decisions made when the feature set was much
| more limited. However, it is somewhat predictable that open
| source general-purpose software will develop faster than specialized
| things with a smaller user base and I expect that trend to continue.
| If you pushed the data store off onto some other well tested code
| it might be easier to concentrate on the unique features.
Somewhat predictable, true, but it's not always clear what projects
are going to make it and which aren't. When I was starting work on
Ganymede, MSQL and MySQL were the open source databases that looked
like they had the mojo going, and neither were adequate.
| >True as well. That's why Ganymede is being sold as a thing to feed
| >NIS and DNS and LDAP and other things for which there is already
| >well-defined API's and client code.
|
| The problem I see with this is that most of these also define
| network API's for updates (maybe all with the new dynamic DNS
| stuff coming around). If you are feeding static data from
| some other source you have to give up any native update protocols.
Tell me about it. We have been using mainly NIS and DNS in our
environment. With NIS we support rpc.yppasswdd integration with a
proxy, and with DNS there isn't yet any DDNS protocol in wide use, but
it's pretty clear that the general problem of intercepting other
updates is a nasty one.
I envy NT-heads with their all-singing, all-dancing,
you-will-use-this-and-don't-worry ActiveDirectory that Microsoft is
doing for them.
Of course, ActiveDirectory always reminds me of the writing on the
One Ring that Gandalf showed to Bilbo.
| LDAP doesn't even map to LDAP very well, given that there is no
| transactional mode to create an object and it's parent in
| an atomic operation. But, if the data can be stored in a
| dbm file, anything else should be able to handle it.
I'd be very open to any well-formed proposals on how to integrate LDAP
support. ;-)
| >As long as external code can be connected to the Ganymede server to
| >support things like NT domain control and LDAP queries, I'm personally
| >going to have to leave off supporting those things natively in favor
| >of proxies and auxiliary servers fed by Ganymede.
|
| Proxies would be great. I'd even be happy to allocate some IP addresses
| to the proxies if the ports for the real services can't be moved.
| Maybe it would be possible to pass queries through to the native
| services but intercept updates and propagate to everything else
| that needs them.
Yeah. Chris Browne suggested implementing CORBA in the Ganymede
server, which would make all of this a lot easier.
| Les Mikesell
| [EMAIL PROTECTED]
--
===============================================================================
Jonathan Abbey [EMAIL PROTECTED]
Applied Research Laboratories The University of Texas at Austin
===============================================================================
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: Questions on Linux Memory Managment
Reply-To: [EMAIL PROTECTED]
Date: Sat, 12 Jun 1999 15:52:19 GMT
Joel Huselius <[EMAIL PROTECTED]> writes:
>On 8 Jun 1999, J.H.M. Dassen (Ray) wrote:
>> Quite frankly, your questions look a lot more like questions asked
>> by instructors in a practical course assignment than ones that come
>> up in students.
>We regret that this discussion is held on Usenet, but after
>J.H.M. Dassen=B4s accusation we cannot be silent.
[...]
>Such a foul and mean allegation demands a direct response,
>and we expect an excuse for this, unfounded and
>incorrect, attack!
I guess that is why "Memory management i Linux" is listed as one of
the examples at http://www.idt.mdh.se/kurser/ct2510/Assignment.html.
Notice the name of the document....
I hope your "fvrdjupningsuppgift" will give you many deep insights ;-)
Bernie
--
============================================================================
"It's a magical world, Hobbes ol' buddy...
...let's go exploring"
Calvin's final words, on December 31st, 1995
------------------------------
From: D Allinson <[EMAIL PROTECTED]>
Subject: Have you used QNX ?
Date: Sat, 12 Jun 1999 18:18:20 +0100
Reply-To: [EMAIL PROTECTED]
Has anyone switched OS from QNX to Linux ? I'd be interested to hear of
why / how easy.
------------------------------
From: "Mark A. Richman" <[EMAIL PROTECTED]>
Crossposted-To: comp.lang.java.corba,comp.object.corba,comp.os.linux.development.apps
Subject: ORB for Linux w/ Java Binding?
Date: Sat, 12 Jun 1999 13:25:27 -0400
Is there an ORB for Linux w/ Java Binding? Hopefully it's free.
-Mark
------------------------------
From: Medical Electronics Lab <[EMAIL PROTECTED]>
Subject: Re: Testing a kernel, boot from floppy?
Date: Fri, 11 Jun 1999 13:18:38 -0500
Lew Pitcher wrote:
> Sorry Dr. Mike, but I missed the first part of this conversation,
> and chose this spot to jump in to.
>
> Konrad: Why can't you
>
> make zdisk
>
> to make a bootable floppy?
>
I tried that, but it didn't work. I think the floppy has to have
a lot of other stuff on it as well as the (compressed) kernel image.
What I will try is to copy the boot disk I made initially when I
loaded from RedHat CD (that still works) and put the new zImage
on that as the kernel. The problem is that the file size will be
wrong, and I bet lilo will go bonkers trying to load data that
isn't there. So far, I can't get lilo to build a boot sector
correctly, so I'm doing something obviously stupid. Unfortunatly,
I don't know what it is!
Patience, persistence, truth,
Dr. mike
------------------------------
From: George MacDonald <[EMAIL PROTECTED]>
Subject: Re: Need help on /proc/[pid]/*
Date: Fri, 11 Jun 1999 18:30:42 GMT
[EMAIL PROTECTED] wrote:
>
> On Wed, 9 Jun 1999 16:41:02 +0100, "John Burton" <[EMAIL PROTECTED]>
> wrote:
>
> >man proc?
> yeah... this is it!
>
> Thanx again!
Last time I checked the docs were a little out of date, you can actually
look at the kernel code for the details, i.e.
/usr/src/linux/fs/proc/array.c
If you want to see it in use, or want to borrow some code to process the
/proc entries have a look at
http://24.1.97.22/gmd/tps/treepsfm.html
In particular ~/Linux/loadValues.c has some routines that might save you
some time. ... Also see the test program lpstat.
There is also a "portable" library that attempts to provide a common interface
to getting process info(meta-data). It's called libgtop and it's part of the
GNU project.
--
We stand on the shoulders of those giants who coded before.
Build a good layer, stand strong, and prepare for the next wave.
Guide those who come after you, give them your shoulder, lend them your code.
Code well and live! - [EMAIL PROTECTED] (7th Coding Battalion)
------------------------------
Crossposted-To: alt.os.linux,comp.os.linux.advocacy,comp.os.misc,comp.unix.advocacy
Subject: Re: TAO: the ultimate OS
From: [EMAIL PROTECTED] (Matthias Buelow)
Date: 11 Jun 99 18:27:30 GMT
In article <[EMAIL PROTECTED]>,
Donovan Rebbechi <[EMAIL PROTECTED]> wrote:
>> Do you have proof that peer-review is not used in the
>>cathedral model?
>
>It's obvious that *more* peer review takes place with open source.
>The only question is "how much more".
This says nothing. I prefer a smaller group of peers who
know what they're doing than a huge bunch of half-knowledgable
opinion-throwers who do not have the necessary background and
insight into the project.
"Open source" does not mandate "bazaar" model. Some of the
best free projects are more "cathedral" (what a horrible
comparison, I'm still missing the actual link between a
buzzing patched marketplace and a religious building and
software engineering), like all the BSD systems. It's not
what I think Mr. Raymond would call "pure cathedral" since
there is a lot of peer review actually going on but still.
It shows that the "cathedral" model is able to produce rather
high quality products in a fairly well organized and up-to-date
way. There are entirely negative examples for both software
engineering types and I dare say that it is not possible at
all to say in advance which model will suit a certain project
better. Certainly a sound mix is best, as always; but we
do not need Mr. Raymonds odd formalism to name the obvious.
------------------------------
From: [EMAIL PROTECTED] (Neil Schemenauer)
Subject: smbios (DMI) support in Linux?
Date: 12 Jun 1999 17:27:16 GMT
I just bought a new machine which includes support in the BIOS
for measuring the CPU temperature (Award BIOS). I believe the
API is defined by a standard called smbios. Is there any support
in Linux for reading this temperature? If not, how would I go
about adding support? I am an experienced C programmer but am
not a kernel expert.
Neil
------------------------------
From: David Knight <[EMAIL PROTECTED]>
Crossposted-To: alt.uu.comp.os.linux.questions,comp.os.linux.networking
Subject: Re: Linux & Cybercafe
Date: 11 Jun 1999 15:36:47 +0100
[EMAIL PROTECTED] (Moritz Moeller-Herrmann) writes:
> On 10 Jun 1999 14:46:08 GMT, David L. Bilbey <[EMAIL PROTECTED]> wrote:
> >What about looking into a smaller, easier to secure browser? Who said that
> >it _has_ to be Netscape?
>
> Well, iexplore.exe won't cut it. Opera hasn't been releases. KFM lacks support
> for java(script) and lynx lacks almost everything. Anything I forgot for
> browsing?
KFM isn't an option as its also file manager, see any problems arising with
respect to starting programs? Although this could be got around by setting
correct permissions
Oh and you forgot about using emacs as a web browser :-)
> I think write protecting (mount readonly)the whole filesystem apart from the
> ~/.netscape/cache directory would protect you from most attacks. Set the shell
> to some script starting Netscape if not running. Disable everything. Give
> permissions to do zilch. Can't be that hard. Maybe not even the cache and use
> a local proxy?
Stopping the settings being saved may stop it being permanent, but will
netscape allow the settings to be changed just for that session, keeping the
changes in memory?
David
--
The superior man understands what is right;
the inferior man understands what will sell.
-- Confucius
------------------------------
From: Marco Fonseca <[EMAIL PROTECTED]>
Subject: Re: Casio QV-11 on Lava DSerial PCI Port, follow-up
Date: Fri, 11 Jun 1999 22:04:41 -0400
Tjeerd Mulder wrote:
>
> I modified the standard serial driver to support several PCI serial
> cards including the Lava one. I have patches for 2.0.36 and 2.3.3. The
> 2.3.3 patches should work for 2.2.9 as well. I did try to get them in
> 2.0.37 but was too late. Hope to get them into the 2.3. series. Please
> tell me if you want to have the patches.
>
> Marco Fonseca schrieb:
> >
> > Marco Fonseca wrote:
> > >
> > > Does anyone out there know how I can go about getting Linux to recognize
> > > the Lava card (which in Win98 device manager appears in two separate
> > > entries: as a parent card and as COM ports 5 and 6) and then Casio
> > > digital camera attached to it?
> > > Thanks,
> > > Marco Fonseca
> >
>
> --
> ======================================================================
> Tjeerd Mulder ! Email: [EMAIL PROTECTED]
> Siemens PC Systeme RD CM !
> Buergermeister Ulrichstr 100 ! Phone: +49 821 804 3549
> 86199 Augsburg ! Fax : +49 821 804 3628
> ======================================================================
Tjeerd:
Thank you very much for your message and the answer is YES provided the
patch works on kernel 2.2.5 as well. I don't think there's any other way
of going about getting Linux 2.2.5 on RedHat 6.0 to recognize the Lava
DSerial 550 PCI card. So, I'd like to try and isntall the patch
(something I've never done before in all of my 6
months of Linux experience!) and then see if I can get the camara to
work as well.
Thanks,
Marco
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list (and comp.os.linux.development.system) via:
Internet: [EMAIL PROTECTED]
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Development-System Digest
******************************
