Linux-Development-Sys Digest #860, Volume #7 Tue, 16 May 00 06:13:09 EDT
Contents:
Re: Need input on developing a unified configuration program for linux (Mongoose)
Re: Need input on developing a unified configuration program for linux (Mongoose)
Re: Need input on developing a unified configuration program for linux (Todd Knarr)
Re: Need input on developing a unified configuration program for linux (Lew Pitcher)
Re: Linux Driver Development (Arnaud Westenberg)
Re: kernel 2.3.99-pre6 ([EMAIL PROTECTED])
Re: ANSI C & void main() ("Tim Roberts")
Re: inline to the extreme??? (Josef Moellers)
Re: Allocating memory at a specific physical locaation ([EMAIL PROTECTED])
Re: Linux compared to Windows 2000 (Matthew King)
Re: Allocating memory at a specific physical locaation (Mathias Waack)
Re: Windows98 IDE driver screwed up Linux UDMA disk access? (Stefan Taferner)
Re: binary compression -- good or bad? ([EMAIL PROTECTED])
Re: Weird behavior of dual-boot Tyan S1854 ([EMAIL PROTECTED])
----------------------------------------------------------------------------
From: Mongoose <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.development.apps,comp.os.linux.misc,comp.os.linux.setup
Subject: Re: Need input on developing a unified configuration program for linux
Reply-To: [EMAIL PROTECTED]
Date: Tue, 16 May 2000 03:16:26 GMT
On 15 May 2000 21:28:07 -0400, [EMAIL PROTECTED] (Alexander
Viro) wrote:
>>users have to learn. I was going to write my program to allow modules
>>in any language. As long as they can do STDIO. They just output their
>
>Erm... Question: "modules" as in "independent binaries" or as in "something
>that can be linked with your code"? The latter case sucks horribly - it
>excludes shell scripts, to start with. The former... What the heck do you
>mean "can do STDIO"? How does the choice of library matter?
Yes independant binaries. And as for stdio, I was enforcing the idea
that every language people are going to use to edit a config file will
have stdio and can be used for a module.
>
>>html page, the web server can serve it and then send data back to the
>>program, using cgi. It won't matter what language its written in and
>>the user will just have to learn html and cgi, which most people
>>already know.
>
><shudder> Considering the quality of HTML floating around (let alone
>CGI - that's a separate rant) I wouldn't use the word "know". Aside of
>the biblical sense, that is...
Well if they can't figure out decent html, they probably wouldn't be
able to figure out how to write a decent module anyways. =)
------------------------------
From: Mongoose <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.development.apps,comp.os.linux.misc,comp.os.linux.setup
Subject: Re: Need input on developing a unified configuration program for linux
Reply-To: [EMAIL PROTECTED]
Date: Tue, 16 May 2000 03:16:42 GMT
Blah, that's basically what I was going to do.
On Tue, 16 May 2000 00:57:10 GMT, [EMAIL PROTECTED] (Christopher
Browne) wrote:
>Centuries ago, Nostradamus foresaw a time when Mongoose would say:
>> I was going to allow remote administration of the system. I know
>>this creates lots of security problems but none that can't be solved.
>>
>> As for linuxconf... linuxconf is actually the reason why I'm making
>>this program. Linuxconf is pretty messy and disorganized. But the main
>>reason is that linuxconf only allows C++ modules to be added into the
>>program, and they have to be written to work with a new interface
>>users have to learn. I was going to write my program to allow modules
>>in any language. As long as they can do STDIO. They just output their
>>html page, the web server can serve it and then send data back to the
>>program, using cgi. It won't matter what language its written in and
>>the user will just have to learn html and cgi, which most people
>>already know. As for the web server, I was going to write my own mini
>>webserver, if there wasn't one already. Just enough to serve the pages
>>on a different port. This way the user configuring their system
>>wouldn't be required to have a web server installed, and I wouldn't be
>>running through port80 creating and more security issues.
>
>Have you looked at WebMin?
>
><http://www.webmin.com/webmin/>
>
>It looks like it is trying to handle many of the same things you are,
>and has the merit that there are already a boatload of modules to help
>manage different services in a modular manner.
>
>Sounds to me like you should probably look into that, if you want to
>do web-based interactive system reconfiguration.
>
>If you want something scripted, you should look at Cfengine.
> <http://www.iu.hioslo.no/cfengine/>
------------------------------
From: Todd Knarr <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.development.apps,comp.os.linux.misc,comp.os.linux.setup
Subject: Re: Need input on developing a unified configuration program for linux
Date: 16 May 2000 03:23:13 GMT
In comp.os.linux.development.system <[EMAIL PROTECTED]>
Mongoose <[EMAIL PROTECTED]> wrote:
> As for linuxconf... linuxconf is actually the reason why I'm making
> this program. Linuxconf is pretty messy and disorganized. But the main
> reason is that linuxconf only allows C++ modules to be added into the
> program, and they have to be written to work with a new interface
> users have to learn. I was going to write my program to allow modules
> in any language. As long as they can do STDIO. They just output their
> html page, the web server can serve it and then send data back to the
> program, using cgi. It won't matter what language its written in and
The C++-only is a point. However, the rest is precisely what linuxconf
in Web interface mode does. It merely resides on a different port, but
it feeds HTML back to the user's Web browser and the user enters data
via standard forms ( there is no such thing as CGI between the browser
and the server ).
> Which kinda brings me to another question. Is passing data through
> stdio a good idea? Won't certain characters, like high ascii
> characters, get lost if you try to pipe them through stdio?
In general Unix streams ( which is what stdin/stdout/stderr are ) are
8-bit clean. The Web server and browser may not like characters not
legal in HTML documents, if any, but that's about it. All you will
need to do, if writing a CGI script, is follow the CGI I/O spec.
--
Collin was right. Never give a virus a missile launcher.
-- Erk, Reality Check #8
------------------------------
From: Lew Pitcher <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.linux,comp.os.linux.development,comp.os.linux.development.apps,comp.os.linux.misc,comp.os.linux.setup
Subject: Re: Need input on developing a unified configuration program for linux
Date: Tue, 16 May 2000 04:10:40 GMT
Mongoose wrote:
>
> Hello,
> I'm am currently developing a linux configuration program that
> will be designed to run on all distros for a college research project.
A broad range, to be sure.
> It is to allow people to create modules for configuring different
> aspects of a linux os. Now these modules can be written in any
> language that can be executed on a system and use stdio. Now to use
> these modules I was thinking of having the modules create cgi forms
> using html.
Well, that eliminates the tomsrtbt distribution.
> This would require the user to have a loopback connect and
> a web browser on their system, two very common and simple items for
> linux.
but doesn't exclude the Linux Router Project distro.
> The webrowser would load the html page, and send a cgi string
> back to the module and the module would edit the system based on the
> users input in the cgi string.
Better ensure that it works with the Lynx characterbased web browser,
otherwise you've eliminated _all_ the small non-X distributions.
> Now I was wondering if most people
> agree with the idea of using a web browser for the system
> administration?
Add in SSL security or something as good to ensure that the lusers
don't reconfigure my system, or that some craker d00d doesn't open up
my system to his pals, and maybe I'll go for it.
> I figure using html is a common way of implementing an
> interface that most people know. They wouldn't have to learn anything
> new in order to create a module for whatever they wish. Plus
> webbrowsers are available in almost every platform possible so
> transportability would be no issue. What does everyone else think of
> this idea?
It's OK, but not 'universal'. It'll work for most commercial
distributions (i.e. the big market ones), but will not be very usefull
for the small distributions.
> Any suggestions?
Stick with a character-based command-line tool for the minimum system.
You can embellish it with shell scripts or perhaps some scripting
language (TCL or Perl or Python or Java) control, but it should be
something that a bare-bones system can run. It should not depend on
network connectivity, advanced GUI functions, or elaborate
configuration databases (which might not be available or might break).
Stay as far away as possible from the "single point of failure". Make
the user interface reasonable and the programming interface dead
simple. Provide lots of documentation. Don't assume anything but the
basic tools. Build it on what ever platform you choose, but _test_ it
on things like LRP or tomsrtbt. When it can run properly on a
nano-linux, then it will be a usefull general purpose tool.
Otherwise, restrict your platform to some common subset (i.e. RedHat
and derivitives), and make it as elaborate as can be done on that
platform. But be aware that you can't do both.
--
Lew Pitcher
Master Codewright and JOAT-in-training
------------------------------
Date: Tue, 16 May 2000 06:54:00 +0200
From: Arnaud Westenberg <[EMAIL PROTECTED]>
Subject: Re: Linux Driver Development
Ben Lull wrote:
> I have recently been getting interested in developping drivers for some
> scanners which I've obtained. I'm not sure where to start. I don't see
> much on the web for driver development. It would be appreciated if
> someone could either post a tutorial on developing linux drivers in C or
> some links where I can get information on developing drivers..
Hi Ben,
As someone noted in an earlier thread, the scanner drivers can be found
through the SANE (Scanner Access Now Easy) project, however SANE itself
isn't a driver. http://panda.mostang.com/sane/
For general information check out the LDP guides. Especially the kernel,
kernel module programming and kernel hackers guide.
You probably wan't to buy the book "Linux device drivers" by Alessandro
Rubini ISBN: 1-56592-292-1
Arnaud
------------------------------
Crossposted-To: uk.comp.os.linux
From: [EMAIL PROTECTED]
Subject: Re: kernel 2.3.99-pre6
Date: Tue, 16 May 2000 05:55:05 GMT
On 15 May 2000, bill davidsen wrote:
>
> In article <[EMAIL PROTECTED]>,
> Charles Blackburn <[EMAIL PROTECTED]> wrote:
>
> | seeing as kernels 2.3.99-pre6 doesn't support ipchains, how do I go about
> | setting up a firewall? any help would be appreciated.
>
> Use iptables instead. I hope to post an errata to the man pages later
> today, unless something else breaks.
Is there not a ipchains (and ipfwadm) module which will allow you to still
use ipchains? I played with masquerading a bit in 2.3.99-pre3 and I am
sure I saw some options to enable ipchains compatibility - never tried it
though, just used iptables instead, so I cant say if it worked or not.
--
Adam
[EMAIL PROTECTED]
[EMAIL PROTECTED]
[EMAIL PROTECTED]
------------------------------
From: "Tim Roberts" <[EMAIL PROTECTED]>
Subject: Re: ANSI C & void main()
Date: Mon, 15 May 2000 23:39:55 -0800
In article <eIKT4.63898$[EMAIL PROTECTED]>, "Mark
Graybill" <[EMAIL PROTECTED]> wrote:
> Look at the sentence again:
>
> "It shall have a return type of type int, but otherwise its type is
> implementation�defined. All implementations shall allow both of the
> following definitions of main: int main() { /* ...
> */ } and int main(int argc, char* argv[]) { /* ... */ }"
>
> The but otherwise provides the exception to the return type of int.
This is the key point where your logic goes awry. The "but otherwise"
says that any part of the declaration of main OTHER THAN THE RETURN
TYPE can be implementation-defined. The return type is strictly
defined and non-negotiable: the return type must be int. Everything
else is open to alteration, but the return type must be int.
Perhaps it is the overloading of the word "type" in the standard that
is confusing you. The "type" of a function includes its return type,
its linkage, and its arguments. Out of that list, everything about
the type of main is implementation-defined EXCEPT the return type, which
must be int. THAT'S what the "but otherwise" says.
> It does not prohibit non-int types.
Yes, it does. "It shall have a return type of type int". That seems
unambiguous to me.
Perhaps this would be more clear if we expanded the word "otherwise"
in this way, which says the exact same thing:
It shall have a return type of type int, but OTHER THAN THE RETURN
TYPE its type is implementation-defined.
> I fail to understand why you cannot see this.
Ditto, I'm afraid.
> If I say the sky is blue, but otherwise it is a different color, doesn't
> say it can only be blue.
Your analogy is flawed. A better analogy is "Your pants must be red, but
otherwise your style of dress is implementation-defined." If the meaning
were truly as you say, then the phrase "It shall have a return type of int"
is completely useless. There is very little useless text in the standard.
--
- Tim Roberts, [EMAIL PROTECTED]
Providenza & Boekelheide, Inc.
------------------------------
From: Josef Moellers <[EMAIL PROTECTED]>
Subject: Re: inline to the extreme???
Date: Tue, 16 May 2000 08:49:42 +0200
Eric Taylor wrote:
> Is there a simple way to turn off all that
> in-lining to see what the difference might be?
> Just removing the word inline would only affect
> a single file.c as many of these inlined procedures
> are in .h files and so get repeated many places.
Have you tried something like adding a -Dinline=3D"" option to the topmos=
t
kernel Makefile?
> Please don't flame me if this is a sensitive issue. I
> was just wondering...
The question seems sensible enough to me.
-- =
Josef M=F6llers
Fujitsu Siemens Computers
SHV Server DS 1
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: Allocating memory at a specific physical locaation
Date: Tue, 16 May 2000 07:37:15 GMT
In article <[EMAIL PROTECTED]>,
Mathias Waack <[EMAIL PROTECTED]> wrote:
>
> Use the "mem=x" command line switch for your kernel, to reserve some
> space for you. After that you can use the remap_page_range function
> to map the space. So you can use only the highest part of memory. If
you
> need other parts, write you own nopage-handler which maps this pages.
>
> --
> Mathias Waack | [EMAIL PROTECTED]
> Tel.: +49 621 181 2717 Fax.: +49 621 181 2713
>
I'm a little confused about 'mem='. When I reserved some upper mem by
'mem=xx' from kernel usage,how can I give its base address to my driver
DMA? And can I map it to a virtual address? Please help me, I'm not very
clear on kernel page.
Thanks in advance!
pubb
Sent via Deja.com http://www.deja.com/
Before you buy.
------------------------------
From: Matthew King <[EMAIL PROTECTED]>
Subject: Re: Linux compared to Windows 2000
Crossposted-To: comp.os.linux,comp.os.linux.development.apps,comp.os.linux.misc
Date: Tue, 16 May 2000 08:42:48 GMT
In comp.os.linux.development.apps Amit Cohen <[EMAIL PROTECTED]> wrote:
> I am looking for a good survey, concerning Linux (mainly RedHat) VS.
> Win2000.
Why mainly RH? RH is not considered `THE' linux (at least not by me ;)
If you're more interested in easy and quality of sysadmin, suse may be a
good comparison. If technical ability of the system, DIY or slackware (i'm
told) would be more interesting
But have fun, this kind of thing has never been done before :-)
Matthew King
--
/------------------------------------------------------------------\
| http://i.am/thenut/ | |
| Finger me for my pgp key or | A king's castle is his home. |
| find it there | |
\------------------------------------------------------------------/
------------------------------
From: Mathias Waack <[EMAIL PROTECTED]>
Subject: Re: Allocating memory at a specific physical locaation
Date: Tue, 16 May 2000 10:56:32 +0200
Hi pubb,
[EMAIL PROTECTED] wrote:
> I'm a little confused about 'mem='. When I reserved some upper mem by
> 'mem=xx' from kernel usage,how can I give its base address to my driver
> DMA?
To use the reserved memory starting at physikal address 'start' with
size 'size' use:
mem = ioremap(start,size);
> And can I map it to a virtual address?
Of course: write your own mmap-methode:
static int my_mmap(struct file* filp, struct vm_area_struct* vma)
{
// some checks for protection, size aso...
if(remap_page_range(vma->vm_start, start,
size, vma->vm_page_prot)) {
return -ENXIO;
}
return 0;
}
The function signatures and names depend on you kernel version,
look at the sources or/and read Rubini.
Mathias
--
Mathias Waack | [EMAIL PROTECTED]
Tel.: +49 621 181 2717 Fax.: +49 621 181 2713
------------------------------
Subject: Re: Windows98 IDE driver screwed up Linux UDMA disk access?
From: Stefan Taferner <[EMAIL PROTECTED]>
Date: 16 May 2000 11:01:31 +0200
[EMAIL PROTECTED] (Paul D. Smith) writes:
> just disabled that weird third IDE controller in the Windows hardware
> profile (Windows won't let me delete it) and Windows then booted again.
Windows hardware profile does not change your bios or hardware.
So you have another problem. Did you maybe change any bios settings
too (before you disabled udma) ?
> But, Linux still hangs, just like before.
>
> I finally went into the BIOS and changed the UDMA support on the first
> IDE controller from what it was ("AUTO") to "OFF".
>
> Now, Linux boots. It also mounts /c, and I can examine files there just
> fine, as before.
>
> However, Linux mounts /c as read-only (even though that's not on my
> mount options) and won't let me write to the disk. If I force a
> re-mount in read-write mode, it works for about 30 seconds, then it
> seems to reset to read-only mode again, on its own!
tail /var/log/messages
There you should find some tips why it fell back into read-only.
Probably because the filesystem is corrupted.
You can install dosfsck and fsck the windows partition, or do
a file system check of the partition in Windows.
-- Stefan
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: binary compression -- good or bad?
Date: Tue, 16 May 2000 09:22:48 GMT
On Tue, 16 May 2000 00:57:21 GMT Christopher Browne <[EMAIL PROTECTED]> wrote:
| If I have fifteen instances of /bin/bash running, which is quite
| possible, if they are compressed, they will all wind up expanding into
| their own independent memory spaces. Whereas if they're left
| uncompressed, they can "demand page" in as needed, which _likely_
| results in substantial sharing of pages between those 15 instances.
If ... and that's a big if, I think ... the system can be made to
recognize the uncompressed "shadow" of the otherwise compressed
object, and establish sharing over the uncompressed copy (which has
to start in RAM but can be swapped out to swap space), then it would
indeed be possible to share it. To do that, the uncompression would
have to all take place at once and then the copy-on-write flag cleared
for the output space. It would be entered in a table of "uncompressed
shadow objects" which would be accessed in place of the original.
The potential benefit _might_ be realized for rarely used, but large
and/or numerous, executeables (and maybe even libraries). OTOH, how
many such programs do you really need?
| Methinks that this compression would make the system a _WHOPPING LOT_
| less efficient...
Not to mention slower :)
While it has potential usefulness in rare and special cases which could
benefit from it, for normal modern computers, this would not be the case.
For example, if disk space is the issue, trim your system of fat and don't
install anything you don't need. If disk space is an extreme issue, you
won't have the space swap space, anyway. It's certainly not something
that would be effective on a ZIP disk based system. And for floppies,
the current practice of compressing initrd and uncompressing it into RAM
is the most optimal, I'm sure you'd agree.
--
| Phil Howard - KA9WGN | My current boycotts: Amazon.Com, DVDs, Mattel, Sony
| [EMAIL PROTECTED] +----------------------------------------------------
| Dallas - Texas - USA | My current websites: linuxhomepage.com, ham.org
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To: alt.comp.periphs.mainboard.tyan
Subject: Re: Weird behavior of dual-boot Tyan S1854
Date: Tue, 16 May 2000 09:32:32 GMT
In comp.os.linux.development.system Jerry Natowitz <[EMAIL PROTECTED]> wrote:
| I just diagnosed a really strange behavior I found on my dual (Linux and
| Win98SE) boot Tyan S1854. I found that if I booted Linux first and then
| tried to boot Windoze, the syetem just sat there doing nothing. Hitting the
| reset didn't change thing, the system still hung right after the message
| about verifying the configuration. A power-cycle would fix things.
|
| After some investigation I found out that my use of modules in Linux for
| the CD-ROMs and Zip drive were the cause. Change:
| CONFIG_BLK_DEV_IDECD=m
| CONFIG_BLK_DEV_IDEFLOPPY=m
| to
| CONFIG_BLK_DEV_IDECD=y
| CONFIG_BLK_DEV_IDEFLOPPY=y
| and all is well. Kernel versions, BTW, are both 2.2.15 and 2.3.99pre9-1
In the case where you had modules, did the hang result when you actually did
load the modules, or when you actually did not load them? Did you try it
both ways? If it fails both ways vs. fails only if loaded or only if not
loaded, that could mean something (although I don't know what that would be).
--
| Phil Howard - KA9WGN | My current boycotts: Amazon.Com, DVDs, Mattel, Sony
| [EMAIL PROTECTED] +----------------------------------------------------
| Dallas - Texas - USA | My current websites: linuxhomepage.com, ham.org
------------------------------
** 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
******************************
