Linux-Development-Sys Digest #330, Volume #6 Mon, 25 Jan 99 23:13:59 EST
Contents:
Re: TAO: the ultimate OS (Tom Harrington)
Re: Configuring system to have multiple ethernet addresses ([EMAIL PROTECTED])
Consultant needed for RH installations... (Graham TerMarsch)
Re: How to tell current kernel config? (Rob Funk)
Re: Kernel 2.2.0pre8 breaks eject-1.5 CDROM (Andrea Borgia)
Re: Secuity hole with perl (suidperl) and nosuid mounts on Linux (Peter Samuelson)
Re: Future Modem support. (Johan Kullstam)
Re: disheartened gnome developer (steve mcadams)
display issues/questions (dman77)
Re: How can I build a Linux system from scratch - NO distribution? (Emile van Bergen)
Re: writing to memory mapped regions---video buffers (Phil Howard)
Re: Modest next goal for Linux ("John De Hoog")
Re: Modest next goal for Linux (Steve Peltz)
Re: Modest next goal for Linux (jedi)
Re: Modest next goal for Linux (Christopher B. Browne)
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (Tom Harrington)
Crossposted-To:
alt.os.linux,comp.os.linux.advocacy,comp.os.misc,comp.sys.misc,comp.unix.advocacy
Subject: Re: TAO: the ultimate OS
Date: 25 Jan 1999 20:25:02 GMT
Reply-To: [EMAIL PROTECTED]
jdn ([EMAIL PROTECTED]) wrote:
: Aren't "Tao" and "Taos" two different things?
Yes. "Taos" is the name of a city in New Mexico, USA. As far as I'm aware,
it is _not_ intended to be a pluralized form of "Tao". Otherwise this thread
has some nasty implications for the Taos tourist industry, i.e. "The Taos
that can be visited is not the true Taos". What it does mean I'm not sure...
--
Tom Harrington --------- [EMAIL PROTECTED] -------- http://rainbow.rmii.com/~tph
"Its 'freedom of speech', as long as you don't say too much"
-The Neville Brothers
Cookie's Revenge: ftp://ftp.rmi.net/pub2/tph/cookie/cookies-revenge.sit.hqx
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To:
comp.dcom.lans.ethernet,comp.os.linux.networking,comp.os.ms-windows.programmer.networks
Subject: Re: Configuring system to have multiple ethernet addresses
Date: Mon, 25 Jan 1999 20:51:14 GMT
If what you are trying to do is to have two cards with an unique ip and
connected to same network to do different tasks, it wont work. as for
microsfot design, multiple computers with same computer name situation
is not allowed.
If you just want your configured connected to multiple networks, you
can add interfaces as much as resource allows.
In article <[EMAIL PROTECTED]>,
Amey Laud <[EMAIL PROTECTED]> wrote:
> I am building a distributed number crunching system, in which a lot of
> data needs to be moved between machines.
> In order to streamline the process and avoid bottlenecks due to
> network/IO latencies, I am considering
> using separate networks to handle the input and the output, that is,
> each system reads from a different physical network
> and writes into a different physical network. This would mean that each
> machine have two IP addresses
> that are configured on separate ethernet cards and can be addressed and
> used explicitly.
> 1. Is such an arrangement possible? (That is, OS and IP support)
> 2. Are there existing examples of such a setup?
> 3. The arrangement might involve heterogenous platforms.
> I am interested specifically on the possibility of such a setup on
> NT/Linux running on Intel (Xeon)/Alpha.
>
> 4. Are there existing message passing API's (such as MPI) based on
> TCP/IP that support such a configuration.
>
> I would glad to get any suggestions in this regard.
>
> Amey Laud.
>
> [EMAIL PROTECTED]
>
> Research Engineer,
> Centre for Signal Processing,
> Singapore 639798
>
>
============= Posted via Deja News, The Discussion Network ============
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
------------------------------
From: Graham TerMarsch <[EMAIL PROTECTED]>
Crossposted-To: linux.redhat.development,comp.os.linux.development,comp.os.linux.help
Subject: Consultant needed for RH installations...
Date: Mon, 25 Jan 1999 22:30:56 GMT
East Coast Client Seeking RH 5.2 Installation Expert
One of my good clients is seeking a consultant to customize the
installation of a large number of RedHat 5.2 Systems in a heterogenous
computing environment. The client needs someone familiar with all
aspects of RH 5.2 installation, from customizing boot, supplemental and
rescue diskettes to probing for hardware to customizing the installation
steps. Location is not a factor and travel will likely not be
required. The opportunity may lead to on-going consulting work.
If you have these qualifications and are interested, please send me your
resume or curriculum vitae to me ([EMAIL PROTECTED]).
--
Graham TerMarsch
------------------------------
From: [EMAIL PROTECTED] (Rob Funk)
Subject: Re: How to tell current kernel config?
Date: 25 Jan 1999 22:58:31 GMT
In article <[EMAIL PROTECTED]>,
Kenneth Porter <[EMAIL PROTECTED]> wrote:
>This is good if the kernel was built locally. My problem is slightly
>different. I downloaded a pre-built kernel from American Megatrends
>(ftp://www.megatrends.com) to support a system running on one of their
>RAID controllers. I'd now like to build a tuned kernel but using theirs
>as a template. Alas, no .config was included with their patch. Is there
>any way to ask the kernel what it has?
If you're lucky, you got kernel include files along with the
kernel.... check for /usr/include/linux/autoconf.h.
--
============== R o b F u n k ==============|========> [EMAIL PROTECTED] <========
"...his hands were doing what they liked to | [EMAIL PROTECTED]
do best -- replacing men like himself with | [EMAIL PROTECTED]
machines." -Vonnegut, "Player Piano" |http://er4www.eng.ohio-state.edu/~funkr
------------------------------
From: Andrea Borgia <[EMAIL PROTECTED]>
Subject: Re: Kernel 2.2.0pre8 breaks eject-1.5 CDROM
Date: Mon, 25 Jan 1999 20:04:46 GMT
On Thu, 21 Jan 1999 14:07:14 +0100,
in article <[EMAIL PROTECTED]>,
Walter van der Schee <[EMAIL PROTECTED]> wrote:
>access, but since 2.2.0pre8 eject says :
>"eject CDROMEJECT ioctl failed for "/dev/hdc" : Operation not Supported"
>Will this be fixed?
Did you try v2.0 of eject? It has been released just a few days ago...
Hope this helps.
--
Alias: [EMAIL PROTECTED]
sysadm on cantina.students.cs.unibo.it
ftp and mirror administrator on ftp.students.cs.unibo.it
Homepage: http://caristudenti.students.cs.unibo.it/~borgia/
------------------------------
From: [EMAIL PROTECTED] (Peter Samuelson)
Crossposted-To: comp.os.linux.misc,comp.lang.perl.misc
Subject: Re: Secuity hole with perl (suidperl) and nosuid mounts on Linux
Date: 25 Jan 1999 17:41:25 -0600
Reply-To: Peter Samuelson <[EMAIL PROTECTED]>
[<[EMAIL PROTECTED]>]
> > It is an matter of opinion (or perhaps even religion) if the
> > user-space hack (a suid-interpreter) for supporting secure suid
> > scripts is considered more or less ugly than the kernel-space hack
> > (using /dev/fd).
> > Personally I think the kernel space hack is (slightly) more ugly.
[Ilya Zakharevich <[EMAIL PROTECTED]>]
> I see: just fix *all* the executables instead of fixing one place.
> Then fix *all* the executable *again* when a security hole with the
> previous fix (as one with perl) is discovered.
Well, I'm sure you know the history.... Many scripting languages are
quite unsuitable for setuid things anyway (Bourne and C shells come to
mind) for many reasons beyond the symlink security hole. Languages
such as Perl, in which it really does make more sense to program
setuid-type stuff, came about comparitively recently. Any scripting
language that puts forth a claim of being as secure as C can just as
well provide its own suidperl-ish hack.
I don't like it, though. I think the user-space hack *is* ugly,
possibly uglier than the kernel-space hack would be. In any case, I
doubt it would be difficult to hack the /dev/fd (or more precisely
/proc/self/fd) feature into Linux. If anyone wants to make the world a
better and arguably safer place, go ahead and write the patch....
> Oh, I forgot that that Linux kernel development model is as (insert
> your favorite word for fascistic) as Emacs :-(.
If you say so. I haven't had any personal experience with the FSF
Emacs developers but I know their reputation. Linus most of the time
seems more than reasonable and open in his development process -- for
small changes he will literally take patches from anybody, and for
large changes he discusses the issues on the mailing lists. And he
releases the alphas / betas plenty often enough to make you feel a part
of things.
--
Peter Samuelson
<sampo.creighton.edu!psamuels>
------------------------------
From: Johan Kullstam <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.hardware
Subject: Re: Future Modem support.
Date: 25 Jan 1999 17:28:09 -0500
"Jason Hardman" <[EMAIL PROTECTED]> writes:
> Hello,
>
> Does anyone out there know if there is any planned support, or someone
> working on the drivers for internal PCI type modems?
you'll have to ask the manufacturer for that one. once they begin
releasing specifications for it, i am sure many linux coders will step
up and write a driver for it. i suggest you don't hold your breath;
you will probably be waiting a long time.
> Unfortunately I bought a Diamond SupraExpress i56k PRO before getting Linux.
> While it isn't totaly necessary for me as I can still use Win98 to do my
> downloading, It is a Royal pain in the butt.
buy a new modem. it is the only way.
iirc multitech makes a usable (i.e., non-winmodem) pci modem. you may
also try an external.
--
johan kullstam
------------------------------
From: [EMAIL PROTECTED] (steve mcadams)
Crossposted-To: comp.os.linux.advocacy,comp.os.linux.development.apps,comp.os.linux.x
Subject: Re: disheartened gnome developer
Date: Mon, 25 Jan 1999 23:31:37 GMT
[Snipped for brevity, quoted material marked with ">"]
On 25 Jan 1999 00:07:23 -0800, Michael Powe
<[EMAIL PROTECTED]> wrote:
>for bank services, in order to drive them away. One banker says, "we
>don't want losers who keep less than $1000 in their accounts and pay
>off their credit cards on time every month."
Cool, those idiots will go broke in no time because it's not the
losers who keep less than $1000 in their (non-interest-bearing)
accounts and pay off their credit cards monthly.
HOWEVER, I agree with a couple of earlier posters that this thread has
far exceeded any connection with the original topic and should either
get back on topic or move to a politics-ng. fwiw. -steve
========================================================
so what? - http://www.codetools.com/showcase
------------------------------
From: [EMAIL PROTECTED] (dman77)
Subject: display issues/questions
Date: Tue, 26 Jan 1999 04:16:42 GMT
Hello all,
At school, I am using a WinNT machine that i will be adding
linux to soon. The trouble is the monitor. It is old and not
multisyncing, and whenever DOS or something just text is viewed, it
screws up the whole image. It will only work properly at a certain
frequency and w/ a certain amount of colors. Is there anyway regular
linux (non X) can set this?
thanks for any thoughts.
------------------------------
From: Emile van Bergen <[EMAIL PROTECTED]>
Subject: Re: How can I build a Linux system from scratch - NO distribution?
Date: Tue, 26 Jan 1999 01:14:46 +0100
Richard Jones wrote:
>
> My real problem is that I only want to refuse
> queries when the dial-up link is down. When the
> link is up (it's started by hand), I would like
> everything to work as normal? I'd really like
> named to look for the existence of a file before
> it tries to forward a query (and if the file isn't
> there, use its own cache exclusively, or else
> reject the request). I suspect that one fine day
> I will end up hacking the functionality into named
> myself ...
I have a setup with some dynamic firewall rules (man ipfwadm) in
addition to the static rules I have in other scripts. I have put those
dynamic ones in a script called fw_dynrules which can be run as
fw_dynrules ipppX { open | close }. Now, if it's closed, it blocks pings
and UDP port 53 (DNS!). There is dial-on demand (ISDN, otherwise try
diald), and it will trigger on output (of course, only to non-blocked
ports (i.e. telnet, ftp, www, etc)) through that interface.
The nifty thing is to have a semaphore per interface. This makes it
possible for a master open/close script to call fw_dynrules ippp0 close
if it's decremented from 1 to 0, and call fw_dynrules ippp0 open if it's
incremented from 0 to 1.
Now, /etc/ppp/ip-up can increment it, while /etc/ppp/ip-down can
decrement it. Also, you can have two icons sitting on your fvwm button
bar, one which decrements the semaphore for the internet interface, and
one which increments it. Think about it; something like this does all I
which think you need.
B.T.W., to block something outgoing, I use the 'reject' option to
ipfwadm: it replies to the sender with an icmp host unreachable message,
so that you won't be bothered by timeouts; as opposed to 'deny'.
Now, go read the man pages!
--
M.vr.gr. / Best regards,
Emile van Bergen (e-mail address: [EMAIL PROTECTED])
This e-mail message is 100% electronically degradeable and produced
on a GNU/Linux system.
------------------------------
From: [EMAIL PROTECTED] (Phil Howard)
Subject: Re: writing to memory mapped regions---video buffers
Date: Mon, 25 Jan 1999 04:36:39 GMT
On 25 Jan 1999 01:37:41 GMT nexus ([EMAIL PROTECTED]) wrote:
| ok how would i write to the memory mapped region of an PCI/ISA video
| card when i get the address for the memory region of a specific card
| from /proc/pci?.....ive tried memcpy, but it gave me a segmentation
| fault.....i want to be able to write to a secondary ISA or PCI video
| card so i can use two monitors.......ok second question...is there a
| video buffer in linux- like a 0xA0000000 or something?....how does linux
| go about throwing stuff on the screen--does it write to the memory
| mapped regions of cards?......thanx
I believe you can memory map /dev/mem or /dev/kmem to access the real
address space from within a process (with appropriate permission).
--
-- *-----------------------------* Phil Howard KA9WGN * --
-- | Inturnet, Inc. | Director of Internet Services | --
-- | Business Internet Solutions | eng at intur.net | --
-- *-----------------------------* philh at intur.net * --
------------------------------
From: "John De Hoog" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.advocacy
Subject: Re: Modest next goal for Linux
Date: Tue, 26 Jan 1999 11:52:17 +0900
Steve Peltz <[EMAIL PROTECTED]> wrote...
>>finally, with the 2.2 kernel, Linux is getting file name caching, which
>>NT has had all along. Those long waits when viewing file lists using a
>
>"file name caching"? The only file names the kernel deals with are in
>directories, which have always been cached just like any other file.
Then what does this mean in a list of 2.2 features?
"Faster file access. Version 2.2 can store filenames in a high-speed cache
in memory, meaning that users won't have to wait for the computer to
retrieve the information off relatively slow hard disks."
OK, maybe CNet news isn't the right place to be finding out this stuff.<g>
But what is this slowness I notice whenever trying to get a long list of
files to display on Linux?
--
John De Hoog
http://dehoog.org
------------------------------
From: [EMAIL PROTECTED] (Steve Peltz)
Crossposted-To: comp.os.linux.advocacy
Subject: Re: Modest next goal for Linux
Date: 26 Jan 1999 01:22:00 GMT
In article <[EMAIL PROTECTED]>,
John De Hoog <[EMAIL PROTECTED]> wrote:
>finally, with the 2.2 kernel, Linux is getting file name caching, which
>NT has had all along. Those long waits when viewing file lists using a
"file name caching"? The only file names the kernel deals with are in
directories, which have always been cached just like any other file.
------------------------------
From: [EMAIL PROTECTED] (jedi)
Crossposted-To: comp.os.linux.advocacy
Subject: Re: Modest next goal for Linux
Date: Mon, 25 Jan 1999 17:00:29 -0800
On Mon, 25 Jan 1999 04:36:56 GMT, John De Hoog <[EMAIL PROTECTED]> wrote:
>[EMAIL PROTECTED] (Andreas Heiss), writing in comp.os.linux.advocacy
>and possibly other newsgroups, commented:
>
>>But if a 300MHz PII NT box is basically unusable during scanning
>>an image with a SCSI scanner, how's that ? (Move the mouse, the
>>mouse cursor moves 2 minutes later ...)
>>Since I have no own scanner, I recently used the NT box of a friend
>>of mine and I couldn't believe how crappy this NT is. In our institute,
>>we have a scanner on a Linux box and the CPU does almost nothing
>>during scanning.
>
>You're lucky you could find a scanner that works under Linux. Mine is not supported,
If he didn't buy the first, cheapest, parallel port scanner
he could find, that likely explains why he had so little
trouble. Hardware that gets little respect within the
community is always going to be problematic.
>so I have no first-hand corroboration for your report. However, I would assume that
>scanning is not the job of NT, but of the SCSI (ASPI) driver and TWAIN driver, which
>are supplied by various third-party vendors. If they are crappy, I'm not sure it
>follows that NT is therefore crappy. Anyway, does Linux have built-in scanner
>support? I thought it was accomplished by add-ons to the Linux system.
Everything except the kernel is an 'add-on' on a Unix system.
Just get the current version of SANE & build it. If your scanner
is on the supported list, you'll be all set after that.
>
>Can you point to a design difference between NT and Linux that would account for the
>speed differences you have observed?
--
Herding Humans ~ Herding Cats
Neither will do a thing unless they really want to, or |||
is coerced to the point where it will scratch your eyes out / | \
as soon as your grip slips.
In search of sane PPP docs? Try http://penguin.lvcm.com
------------------------------
From: [EMAIL PROTECTED] (Christopher B. Browne)
Crossposted-To: comp.os.linux.advocacy
Subject: Re: Modest next goal for Linux
Reply-To: [EMAIL PROTECTED]
Date: Tue, 26 Jan 1999 03:58:03 GMT
On Tue, 26 Jan 1999 11:52:17 +0900, John De Hoog <[EMAIL PROTECTED]> posted:
>Steve Peltz <[EMAIL PROTECTED]> wrote...
>>>finally, with the 2.2 kernel, Linux is getting file name caching, which
>>>NT has had all along. Those long waits when viewing file lists using a
>>
>>"file name caching"? The only file names the kernel deals with are in
>>directories, which have always been cached just like any other file.
>
>Then what does this mean in a list of 2.2 features?
>
>"Faster file access. Version 2.2 can store filenames in a high-speed cache
>in memory, meaning that users won't have to wait for the computer to
>retrieve the information off relatively slow hard disks."
>
>OK, maybe CNet news isn't the right place to be finding out this stuff.<g>
>
>But what is this slowness I notice whenever trying to get a long list of
>files to display on Linux?
The slowness is in *sorting* the filenames. If you have a lot of files, and
don't specify the "-f" option, the OS (or maybe it's some other layer) has
to sort in some form of lexicographical order.
Older versions of Linux stored such information in essentially linear lists,
which has the result that as directories get big, access times to successive
filenames get longer and longer.
Ergo, combine all this, and you get slow directory listings for very large
lists of files.
Newer versions use hashing, if memory serves, which is faster, albeit with
some aspects of unsafety.
Reiserfs, currently in experimental stages, treats the directory structure
as a B-tree, and can do considerably better and keeps a useful ordering of
files.
--
Those who do not understand Unix are condemned to reinvent it, poorly.
-- Henry Spencer <http://www.hex.net/~cbbrowne/lsf.html>
[EMAIL PROTECTED] - "What have you contributed to Linux today?..."
------------------------------
** 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
******************************
