Linux-Misc Digest #363, Volume #21 Wed, 11 Aug 99 11:13:12 EDT
Contents:
Re: Marx vs. Nozick (MK)
Re: Marx vs. Nozick (MK)
Re: Terminalemulation from WinNT to Unix (Knut Erik =?ISO-8859-1?Q?Lang=F8?=)
Upgrade from RH5.2 to 6 ("Aaron Dershem")
Re: lower to upper case? (Coy A Hile)
Re: Red Hat 6.0 Installation problem (Ross Smith)
Re: Installing Staroffice 5.1 on Slackware 4.0 ("Noah Roberts (jik-)")
Re: Red Hat 6.0 Installation problem (Glen Brittle)
Anybody use PostgreSQL with Redhat 6.0 ?? (Glen Brittle)
Re: Tar Limitations (Robin Smith)
Tar Limitations ("Andrew Gilbert")
Re: Connect to ISP when phone rings: is this possible? (Patrick M. Geahan)
Re: Can't log in (Viktor =?iso-8859-1?Q?Djupsj=F6backa?=)
quick Perl Question ("Matthew D. Melbert")
Terminalemulation from WinNT to Unix (Ludger Sicking)
SUMMARY: Re: tar hoses system ([EMAIL PROTECTED])
Re: My Linux box was hacked! (Rene Grothmann)
Re: xterm: "no available ptys" (Steve Gage)
Re: Marx vs. Nozick (Johan Kullstam)
What's The best Firewall/Proxy Software ("Pat Duczyminski")
Re: Linux assembly, etc (Johan Kullstam)
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (MK)
Crossposted-To: comp.os.ms-windows.advocacy,comp.os.linux.advocacy,gnu.misc.discuss
Subject: Re: Marx vs. Nozick
Date: Wed, 11 Aug 1999 12:45:40 GMT
Reply-To: [EMAIL PROTECTED]
On Wed, 11 Aug 1999 13:27:44 +1159, [EMAIL PROTECTED] (Rodger
Donaldson) wrote:
>>UMM, excuse me, but doesn't *every* OS manufacturer want this? Doesn't
>>Linus? (in fact Linus is content with nothing short of World
>>Domination)
>
>It was a joke, son. Linus has gone on record repeatedly as an advocate of
>choice who wants people to be able to walk into a store and get Windows,
>MacOS, Linux, or whatever.
The one is not necessarily contradicting another. Sure OS seller
advocates choice -- otherwise, how could he get people to
choose _his_ OS?
Marcin Krol
==================================================
Reality is something that does not disappear after
you cease believing in it - VALIS, Philip K. Dick
==================================================
Delete _spamspamlovelyspam_ from address to email me
------------------------------
From: [EMAIL PROTECTED] (MK)
Crossposted-To: comp.os.ms-windows.advocacy,comp.os.linux.advocacy,gnu.misc.discuss
Subject: Re: Marx vs. Nozick
Date: Wed, 11 Aug 1999 12:57:17 GMT
Reply-To: [EMAIL PROTECTED]
On 10 Aug 1999 10:45:02 -0400, Johan Kullstam
<[EMAIL PROTECTED]> wrote:
>the problem is that the market is failing to produce optimum results.
>consider the software market. it is extreme in that
>1) most of the cost is non-recuring initial development. (writing the
> software costs, stamping out cd-roms is nearly free.)
>
>2) the value of the product to the user/customer increases as more
> people use it. (sharing files and information is important.)
>a popular product has low cost when ammortized over many
>installations. the value is high when many people use it.
>therefore, once a product becomes dominant it is nearly impossible to
>dislodge.
I don't agree. Windows is not "nearly impossible to dislodge". It's
just BG seems to be the only one who has discovered the formula
for practical, popular OS for open platform. Other guys on the
market were, and still are, about Big OS for Big Computer and
they were not able to do anything else (recommended reading:
"Alice in UNIX land"). MacOS is about closed hardware platform,
and the whole thing tends to be expensive. Till this day the industry
stubbornely refuses to learn even though the dominance of MS
has nearly beaten them into the ground with clue-stick. You could
drive over them in clue-tank, but they still don't have a clue.
All you need to do is to provide alternative that is truly better.
Users quickly get bored of old things. Otherwise, those newspapers
with CDs full of dozens of programs _each month_ would not sell.
>the standard axioms assumed by classical economists do not
>hold for software. the market is failing to produce optimum results.
By no means: it's just the very special situation. Historically
speaking, the very new situation: computer on each desk,
_totally_ different demographics of users, and so on. The popular
solution is completely different from high-end solution. Accountants
wanted to lynch Henry Ford when he proposed so low prices
for his cars. They thought HF was crazy. He wasn't crazy, just
his competition producing high-end expensive cars was put
out of business. They were not able to adapt to new conditions.
Mentally. They had their set of beliefs that they refused to drop.
So they died. In that time, you would claim that "market is
failiing to produce optimum results" as well. It's not: it is
just in phase of starving exceptionally dense and stubborn
producers to death.
>what do you do when the market doesn't work? usually, you invoke the
>government. this has been done for the armed forces, public education
>&c. the government isn't very good at optimizing returns, but
>sometimes there isn't a better alternative.
No, no. This is burning house to roast a chicken.
Marcin Krol
==================================================
Reality is something that does not disappear after
you cease believing in it - VALIS, Philip K. Dick
==================================================
Delete _spamspamlovelyspam_ from address to email me
------------------------------
From: Knut Erik =?ISO-8859-1?Q?Lang=F8?= <[EMAIL PROTECTED]>
Subject: Re: Terminalemulation from WinNT to Unix
Date: 11 Aug 1999 15:34:52 +0200
[Ludger Sicking <[EMAIL PROTECTED]>]
| I want to log into a unix machine from a WinNT machine.
| On the WinNT machine I run a telnet with termtype vt100.
| But there some keyboard parts aren't interpreted in the right way.
| I want to use the Tab and the up, down and left and right keys.
| I only get the Control abbreviation. How can I configure my program
| to interpret the keys in the right way ?
Use another telnet-application than the default windows-one. E.g. Crt.
------------------------------
From: "Aaron Dershem" <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux,comp.os.linux
Subject: Upgrade from RH5.2 to 6
Date: Wed, 11 Aug 1999 08:26:29 -0500
I've gotten RH5.2 installed at home and it runs just fine. I even upgraded
the kernel to 2.2.6 with no hitches. Are there any compelling reasons to
upgrade to 6.0? If so, is there a lot of work involved, or is it just a
matter of upgrading the RPMs?
Thanks,
Aaron Dershem
------------------------------
From: [EMAIL PROTECTED] (Coy A Hile)
Subject: Re: lower to upper case?
Date: 11 Aug 1999 09:49:24 -0400
In article <[EMAIL PROTECTED]>,
scable <[EMAIL PROTECTED]> wrote:
>scable wrote:
>>
>> In bash, tcsh, or csh, is there a way -- be it a builtin, a utility, or
>> some way of writing a script -- to take an arbitrary string input and
>> turn it into a string whose letters are all upper case?
>
try cat foo | tr a-z A-Z
Coy
--
Coy Hile
[EMAIL PROTECTED]
"Theirs not to reason why; theirs but to do...."
Tennyson, "Charge of the Light Brigade"
------------------------------
From: Ross Smith <[EMAIL PROTECTED]>
Crossposted-To: jaring.os.linux
Subject: Re: Red Hat 6.0 Installation problem
Date: Thu, 12 Aug 1999 02:27:39 +1200
Tan wrote:
>
> But how am I going to create a partition such that it is reside under
> cylinder 1024? Don't the primary dos partition already used up all the
> space below cylinder 1024? I don't sure what is this cylinder stuff.
You can use fdisk (either version) to check how big the cylinders are,
and how many the primary partition is taking up. Typically, 1024
cylinders will be 512 MB, but that varies on some makes of hard drive.
> From what I read in the Linux Installation Manual, if I have 2 HDD (as
> primary and secondary master) and 1 CD-ROM (as primary slave), then
> I've to install my linux one the PRIMARY master HDD, I can't install
> it in the secondary master HDD. Am I interpret it correctly?
That's correct. However, what you can do instead is swap the second HD
with the CD-ROM, so you have HDs on the primary master and slave, and
the CD on the secondary master. Then you can put your Linux boot
partition on (the first 1024 cylinders of) your second HD. Boot
partitions must be on the primary IDE controller, but they can be on the
slave as well as the master. My system is set up that way (Windows boot
on primary master, Linux boot on primary slave).
--
Ross Smith ....................................... Auckland, New Zealand
<mailto:[EMAIL PROTECTED]> ........ <http://crash.ihug.co.nz/~r-smith/>
"For ten years Caesar ruled with an iron hand, then with a wooden
foot, and finally with a piece of string." -- The Goon Show
------------------------------
Subject: Re: Installing Staroffice 5.1 on Slackware 4.0
From: "Noah Roberts (jik-)" <[EMAIL PROTECTED]>
Date: 05 Aug 1999 23:05:01 -0700
[EMAIL PROTECTED] (Mike Trettel) writes:
> I'm trying to install SO 5.1 on slack 4.0, and keep getting a segfault
> when I run the install script. Glibc 2.0.7 is installed properly, as per
> the SO 5.1 directions, so I imagine some other library is out of whack,
> and/or some link somewhere is incorrect. Any pointeres and/or tips would
> be appreciated.
rm /lib/libpthread.so.0
ln -s /lib/libpthread-0.7.so /lib/libpthread.so.0
You will have to do this every time you reboot unless you remove the
other libpthread.
------------------------------
From: Glen Brittle <[EMAIL PROTECTED]>
Crossposted-To: jaring.os.linux
Subject: Re: Red Hat 6.0 Installation problem
Date: Wed, 11 Aug 1999 09:28:46 -0400
This is a multi-part message in MIME format.
==============F4680AD577BC7867828C1D7A
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
I changed my config to Primary master - Windoze , Primary slave to Linux
, Secondary Master CDROM
and loaded lilo into MBR ... works great - no conflicts
Britts
Tan wrote:
>
> OK, for my secondary HDD, now what I have to do is to delete my
> extended dos partition (which is currently unused), but I can keep my
> primary dos partition (which is currently used to store my files).
> Then I've to use Linux's fdisk to partition the HDD. Is these correct?
>
> But how am I going to create a partition such that it is reside under
> cylinder 1024? Don't the primary dos partition already used up all the
> space below cylinder 1024? I don't sure what is this cylinder stuff.
>
> From what I read in the Linux Installation Manual, if I have 2 HDD (as
> primary and secondary master) and 1 CD-ROM (as primary slave), then
> I've to install my linux one the PRIMARY master HDD, I can't install
> it in the secondary master HDD. Am I interpret it correctly?
==============F4680AD577BC7867828C1D7A
Content-Type: text/x-vcard; charset=us-ascii;
name="britts.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Glen Brittle
Content-Disposition: attachment;
filename="britts.vcf"
begin:vcard
n:Brittle;Glen
tel;fax:613 525 3783
tel;work:613 525 2784
x-mozilla-html:FALSE
org:Alexandria Moulding / CanAm Millwork
adr:;;95 Lochiel St E;Alexandria;Ontario;K0C 1A0;Canada
version:2.1
email;internet:[EMAIL PROTECTED]
title:Manager Information Systems
x-mozilla-cpt:;-23312
fn:Glen Brittle
end:vcard
==============F4680AD577BC7867828C1D7A==
------------------------------
From: Glen Brittle <[EMAIL PROTECTED]>
Subject: Anybody use PostgreSQL with Redhat 6.0 ??
Date: Wed, 11 Aug 1999 09:34:18 -0400
This is a multi-part message in MIME format.
==============70608F018E2017166F87899A
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
I am having problems with PostgreSQL installations. :(
I would like to be able to access PostgreSQL from Perl also.
I am a newbie so please be patient.
I have a Pentium 266mmx Linux server. I created it with Redhat 6.0 and
installed everything.
I also tried to install all of the RPM's on the Contributed CD for
PostgreSQL. All went in
except one :
rpm -U -h -v --force postgresql-perl-6.3.2-6.i386.rpm
gave me an error:
error: failed dependencies:
libpq.so.1 is needed by postgresql-perl-6.3.2-6
I could not figure out where this file comes from ...
I went to Redhat website and found updated versions .
I downloaded the following to my pc then ftp'd then to the server in
/tmp:
postgresql-6.5.1-1.i386.rpm
postgresql-perl-6.5.1-1.i386.rpm
postgresql-devel-6.5.1-1.i386.rpm
postgresql-python-6.5.1-1.i386.rpm
postgresql-jdbc-6.5.1-1.i386.rpm
postgresql-server-6.5.1-1.i386.rpm
postgresql-odbc-6.5.1-1.i386.rpm
postgresql-tcl-6.5.1-1.i386.rpm
I killed the postmaster process.
From /tmp I ran:
rpm -Uvh *.rpm
error: failed dependencies:
postgresql-clients conflicts with postgresql-6.5.1-1
HELP ... I am not sure what to do next.
Glen Brittle
I currently have installed on RedHat 6.0 the following:
postgresql-clients-6.4.2-4
postgresql-data-6.4.2-4
postgresql-jdbc-6.3.2-6
postgresql-6.4.2-4
postgresql-devel-6.4.2-4
==============70608F018E2017166F87899A
Content-Type: text/x-vcard; charset=us-ascii;
name="britts.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Glen Brittle
Content-Disposition: attachment;
filename="britts.vcf"
begin:vcard
n:Brittle;Glen
tel;fax:613 525 3783
tel;work:613 525 2784
x-mozilla-html:FALSE
org:Alexandria Moulding / CanAm Millwork
adr:;;95 Lochiel St E;Alexandria;Ontario;K0C 1A0;Canada
version:2.1
email;internet:[EMAIL PROTECTED]
title:Manager Information Systems
x-mozilla-cpt:;-23312
fn:Glen Brittle
end:vcard
==============70608F018E2017166F87899A==
------------------------------
From: Robin Smith <[EMAIL PROTECTED]>
Subject: Re: Tar Limitations
Date: 11 Aug 1999 15:18:18 +0100
"Andrew Gilbert" <[EMAIL PROTECTED]> writes:
> We are having a few problems using tar with large volumes of data. For one,
> we can't create an archive with more than 2147483648 bytes in it
> (suspicously looking like a signed int problem). For two, when trying to
> read large archives (7 Gig +) from a third party we get erratic behavior,
> with tar skipping headers and throwing garbage on the screen (suspicously
> looking like overflow or pointer issues).
>
> We are using Red Hat 6.0. Are there known limitations to tar? Can it only
> handle 2 gig archives? How many items can be in an archive?
>
> This is NOT a disk space problem (unless perhaps in /tmp or /var) as the
> filesystems typically have 10 to 15 GB of free space.
>
> Any help would be appreciated.
>
> Thanks.
Never tested it but Linux should currently have a 2Gb file size limit.
Have you tried named pipes to get around the problem?
e.g.
mkfifo /tmp/pipe ( is this the command to make a named pipe in Linux
or is it mknod? )
split -b 1024m /tmp/pipe split.aaa &
tar cvf /tmp/pipe *
The above was typed in without testing but the idea is to create a
named pipe, start a background process to read information from the
named pipe and create lots of smaller files from the input from the
pipe. A finally send the output of tar to the named pipe.
This is a very useful technique from getting around the 2Gb limit.
HTH
Robin
------------------------------
From: "Andrew Gilbert" <[EMAIL PROTECTED]>
Subject: Tar Limitations
Date: Wed, 11 Aug 1999 09:49:31 -0400
We are having a few problems using tar with large volumes of data. For one,
we can't create an archive with more than 2147483648 bytes in it
(suspicously looking like a signed int problem). For two, when trying to
read large archives (7 Gig +) from a third party we get erratic behavior,
with tar skipping headers and throwing garbage on the screen (suspicously
looking like overflow or pointer issues).
We are using Red Hat 6.0. Are there known limitations to tar? Can it only
handle 2 gig archives? How many items can be in an archive?
This is NOT a disk space problem (unless perhaps in /tmp or /var) as the
filesystems typically have 10 to 15 GB of free space.
Any help would be appreciated.
Thanks.
------------------------------
From: Patrick M. Geahan <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.networking
Subject: Re: Connect to ISP when phone rings: is this possible?
Date: 11 Aug 1999 13:14:23 GMT
In comp.os.linux.misc Glitch <[EMAIL PROTECTED]> wrote:
: how is this possible if u dont know the IP address of your computer once
: its connected to your ISP? Obviously its going to be different each time
: u connect and since u would be telnetting u would need to know the IP of
: the computer, right?
Certain dynamic IP name resolving services will let you update the IP addy
by email or some such, so that could work. In addition, he could simply
have his PPP login script mail him the IP addy after he gets connected.
--
=======Patrick M [EMAIL PROTECTED]=======ICQ:3784715==========
USENET Quote of the Week: "who makes the best Linux? Microsoft?" - JY on
comp.os.linux.misc
------------------------------
From: Viktor =?iso-8859-1?Q?Djupsj=F6backa?= <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux,alt.os.linux.mandrake,comp.os.linux.setup
Subject: Re: Can't log in
Date: Wed, 11 Aug 1999 13:59:32 +0000
Igor Raznatovic wrote:
> For some strange reason, I am receving this message when I try to log
> on:
>
> "Startup program /etc/X11/xdm/GiveConsole exited with non-zero status.
> please contact your system administrator."
>
> I have to add that I have a PC 350mhz with Mandrake on a 3.8Gb
> partition. It boots directly in level 5 so now I cant access prompt
> (frustrating right)...Can someone help me with this?
If you use LILO, type "linux 3" to boot into runlevel 3.
Viktor D.
------------------------------
Date: Wed, 11 Aug 1999 10:42:05 -0400
From: "Matthew D. Melbert" <[EMAIL PROTECTED]>
Reply-To: [EMAIL PROTECTED]
Crossposted-To: alt.os.linux.caldera
Subject: quick Perl Question
I have looked in about EVERY O'Reilly book out there trying to find out
if there is a PERL function that convers an integer (decimal) value to a
Hex value. I would appreciate any help.
TIA
Matthew D. Melbert
------------------------------
From: Ludger Sicking <[EMAIL PROTECTED]>
Subject: Terminalemulation from WinNT to Unix
Date: Wed, 11 Aug 1999 14:32:24 +0200
Hi,
my problem is not Linux specific. The following one:
I want to log into a unix machine from a WinNT machine.
On the WinNT machine I run a telnet with termtype vt100.
But there some keyboard parts aren't interpreted in the right way.
I want to use the Tab and the up, down and left and right keys.
I only get the Control abbreviation. How can I configure my program
to interpret the keys in the right way ?
Special thanks,
Ludger
------------------------------
From: [EMAIL PROTECTED]
Subject: SUMMARY: Re: tar hoses system
Date: Wed, 11 Aug 1999 13:39:30 GMT
I meant that /download is the mount point of the slice /dev/hda6 but
that the slice id (hda1, hda2, etc.) shouldn't matter.
In any case, I have learned that apparently there is a bug in the kernel
version I am running...the disk buffers can become filled, killing
processes. I'll try upgrading the kernel and see if that works.
Thanks to all who responded!
Greg
In article <[EMAIL PROTECTED]>,
Tim Moore <[EMAIL PROTECTED]> wrote:
> > tar cvf /dev/st0 /download
> >
> > where download is a filesystem (/dev/hda6 but that shouldn't matter)
>
> /dev/hda6 is a raw device not a file system. Stick with actual file
> systems when using tar, such as whatever is mounted on /dev/hda6. If
> you want raw device backups try dd(1) or dump(8).
> --
> timothymoore "Everything is permitted. Nothing is forbidden."
> bigfoot WS Burroughs.
> com
>
Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
------------------------------
From: Rene Grothmann <[EMAIL PROTECTED]>
Crossposted-To: comp.security.unix,comp.os.linux.setup
Subject: Re: My Linux box was hacked!
Date: Wed, 11 Aug 1999 16:16:07 +0200
> Those who lack the prowess well beyond that of their "normal" (read:
> "merely competent") peers to create and to innovate have no basis to
> claim to be hackers. Your usage of the word is incorrect and performs
> a disservice to a community to which you should be showing a great
> deal of veneration.
>
> http://www.tuxedo.org/~esr/jargon/html/entry/hacker.html
>
> I cannot help but wonder if my words will have any effect on your
> usage, since the one paragraph I quoted includes at least two
> additional counts of insufficient English mastery:
Sorry, but English is not my native language. So, what does "prowess"
mean? I could not find it in my dictionary. And besides, to which
category of hackers in the cited text are you counting yourself. Maybe
to the crackers?
I too "cannot help but wonder" what kind of personality would ever want
to write such a post here.
Rene.
------------------------------
From: Steve Gage <[EMAIL PROTECTED]>
Subject: Re: xterm: "no available ptys"
Date: Wed, 11 Aug 1999 13:57:31 GMT
"J.H.M. Dassen (Ray)" wrote:
>
> Steve Gage <[EMAIL PROTECTED]> wrote:
> >I finally cleanly installed RH 6.0 after getting tired of trying to morph
> >5.1 any further :-) Alas, in X I can't get an xterm. Upon leaving X, I can
> >see the message:
> >
> >"xterm: no available ptys"
> >
> >Anyone know how to fix this?
>
> RH 6.0 is glibc2.1 based. The combination of glibc2.1 and kernel 2.2.x
> supports Unix98 ptys. It may be a matter of
> - upgrading your xterm package to one that uses Unix98 ptys
> - upgrading your kernel to a 2.2.x one
> - recompiling your kernel to enable Unix98 pty support
> - creating /dev/ptmx and ensuring you have a proper entry for /dev/pts in
> your /etc/fstab.
>
> It's difficult to tell.
It turned out to be my fstab. After installing RH6, I copied my old
fstab back in, in which I had commented out the /dev/pts entry during
some experimentation with a 2.0.x kernel. Thanks much!
- Steve
------------------------------
From: Johan Kullstam <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.advocacy,comp.os.linux.advocacy,gnu.misc.discuss
Subject: Re: Marx vs. Nozick
Date: 11 Aug 1999 09:33:52 -0400
Ketil Z Malde <[EMAIL PROTECTED]> writes:
> [EMAIL PROTECTED] writes:
>
> >> 3) five year plans
>
> > There is something wrong with this? I have a 5 year plan for my own
> > life.
>
> The wrong part is the idea that you can have a plan economy on a large
> scale, i.e. replacing the market forces with detailed control from
> above. What this leads to is that whatever seems important to the
> current leader (e.g. building a giant tractor factory in Minsk, having
> a space program, winning olympic medals) gets a lot of resources,
> while less visible concerns (making washing machines, bicycles, food)
> gets less attention.
>
> Worse, efficiency is no longer a goal, to get resources you must get
> the attention of the resource allocator, or government official, so
> this is what everybody strives for. Even without corruption, this
> isn't going to work very well.
>
> In a capitalistic system, to get resources you need to convince
> investors of your profitability potential or earn money by selling
> products or services, and thus it is in a company's interest (among
> other things) to be competitive by making as good a product as
> possible with as low resource consumption as possible.
yes but large corporations act like communist states in this regard.
projects sponsored by vice presidents with their own agenda often have
no relation to what the customer wants or expects. these are exactly
like the five-year-plans ridiculed in the west.
no one buys stock based on reaping dividends; you buy stock hoping
some other mark will come along and offer you more money for it. i
don't know *why* they (the marks) do this. they just tend to do it.
you play to what works on stock buyers perceptions and developing a
sound business plan isn't nearly as effective a boost on stock prices
as is laying off 10% of your workforce. to be sure, an unprofitable
and bankrupt company will not fetch much of a price, but that is just
because others won't buy your stock. it's still a piece of paper with
only the most tenuous link to the company.
--
johan kullstam
------------------------------
From: "Pat Duczyminski" <[EMAIL PROTECTED]>
Subject: What's The best Firewall/Proxy Software
Date: Wed, 11 Aug 1999 09:11:52 -0400
I was wondering what is the best (free of course) firewall/proxy software to
use on Mandrake to connect 3 other computers (Win98 machines) on my personal
LAN. I uesd Wingate when I was running Win98 but I want to go totally
Linux. A good GUI would be nice, but command line wouldn't be a problem
with good Docs. Any help would be appreciated.
------------------------------
From: Johan Kullstam <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.development.apps,comp.os.linux.development.system
Subject: Re: Linux assembly, etc
Date: 11 Aug 1999 08:54:26 -0400
Josef M�llers <[EMAIL PROTECTED]> writes:
> Alexander Viro wrote:
> > =
>
> > In article <[EMAIL PROTECTED]>,
> > Johan Kullstam <[EMAIL PROTECTED]> wrote:
> > >of course it's not *really* a shared library. however, on the surface=
>
> > >it shares a couple of attributes:
> > >
> > >1) system calls are done by a C subroutine call mechanism. in x86 you=
>
> > > push args onto the stack and use a `call' instruction.
> > =
>
> > Check the facts, please. It's done by putting the values into registers=
>
> > followed by int 0x80. No trace of call. We are *not* using the call
> > gate - it's a trap gate and nothing is copied from the caller's stack.
> > Check arch/i386/kernel/{trap.c,entry.S} for details. Or just do the fol=
> lowing:
> > al@bird:/tmp$ ar x /usr/lib/libc.a write.o
> > al@bird:/tmp$ objdump --disassemble write.o |less
>
> Could it possibly be that you two (Alexander and Johan) happily talk
> past another?
no. alexander is right. i was wrong.
> If I wrote a C program, I would initiate a system call (e.g.
> open/close/read/write) by calling a subroutine/function (possibly) in
> the standard-C-library (g)libc. This is usually but not necessarily a
> shared library. My process would still be in user land, parameters would
> (usually) be passed by pushing them onto the stack and retrieving them
> from there. The passing of arguments and the calling of functions is a
> compiler issue.
> (This is Johan's point of view)
i was thinking that calls to kernel would be patched with the proper
addresses at load time. this is not the case for ia32/x86 linux.
> The library function would then, in turn, do the actual system call,
> e.g. by loading the parameters into registers and executing an INT
> instruction. This would switch into system mode and enter the kernel.
> This is a kernel implementation issue.
> (This is Alexander's statement)
yes.
> Having a library between the user code and the kernel call has the
> advantage that the source code is independent of the kernel-entry
> mechanism: No matter how the transition from user to kernel mode was
> implemented, INT 0x80, CALL gate, SYSCALL instruction, or even
> attempting to execute an illegal instruction, the source code would
> still refer to it as a function. Some "system calls" may even be
> implemented partially or wholly in user land, e.g. a kernel
> implementation may choose to map the proc structure read-only into the
> user address space, getpid(), getuid() and friends would then require no
> entry into kernel mode.
you are correct. the C wrappers of the actual kernel calling
mechanism allow easy porting of the C code. this has been a big win
for unix and C.
> Having a shared library has the additional advantage that a program
> gains a certain degree of binary portability, because the transition
> code is added at runtime and a different library may be linked at
> program start time to reflect the change in calling method.
>
> Hoping to end a senseless debate,
not senseless. i have learned something. i hope others have too.
--
johan kullstam
------------------------------
** 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.misc) 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-Misc Digest
******************************
