Linux-Development-Sys Digest #666, Volume #7 Thu, 9 Mar 00 16:13:19 EST
Contents:
Max number of Telnet ("Darsh")
Re: Binary compatibility: what kind of crack are they smoking? (Peter Samuelson)
Re: select on sockets. ("P.G.Hamer")
Re: Creating VB RAD Style Application to Linux (Peter Samuelson)
Re: Plug and Pray PCI Modem support (Peter Samuelson)
Re: Why a file system ? (Peter Samuelson)
Re: Req. for info on rev. engineering an ISA (Backer) card (Joe Pfeiffer)
ERROR MESSAGE: Unable to load interpreter ("Mark Tranchant")
Re: Recompiled Kernel W/SMP support (Charlie Martin)
Re: memcpy over pci is slow (Alan Donovan)
Re: Can i get a MAC address ? (Alan Donovan)
Re: Impasse with 2 SCSI controllers, kernel mods required? (Stefaan A Eeckels)
Re: Why a file system ? (David Fox)
Re: Tape Driver EOF (Marc SCHAEFER)
Re: Swapping HDs... how best to do it? ("Paul D. Smith")
How to set kernel name? I have (none). (M Sweger)
kmalloc allocate swappable memory? ("Raj Suri")
device or resource busy ERROR (Surya)
init_module: device or resource busy error (Surya)
Re: init_module: device or resource busy error (Surya)
----------------------------------------------------------------------------
From: "Darsh" <[EMAIL PROTECTED]>
Subject: Max number of Telnet
Date: Thu, 9 Mar 2000 14:14:22 +0100
Hi,
Why can't I open more than 5 concurrents telnet sessions ? how can I change
this number ?
thanks !
Darsh
------------------------------
From: [EMAIL PROTECTED] (Peter Samuelson)
Crossposted-To: comp.os.linux.advocacy,comp.os.linux.setup
Subject: Re: Binary compatibility: what kind of crack are they smoking?
Date: 9 Mar 2000 07:26:20 -0600
Reply-To: Peter Samuelson <[EMAIL PROTECTED]>
[Colin Watson <[EMAIL PROTECTED]>]
> It's not clear why the Qt developers would want to give such
> permission if they really believe in the terms of their licence,
> though I suppose money talks. :(
It's their whole business model. Just like Aladdin (distributors of
the original Ghostscript). Free for noncommercial use, as it were.
How do you *expect* Troll Tech to make money? They're not exactly
nonprofit, now.
(Don't anybody read me "The Magic Cauldron". I know. I'm just saying
that TT expects to make their money selling Qt licenses, not that there
isn't *any* other way.)
Peter
------------------------------
From: "P.G.Hamer" <[EMAIL PROTECTED]>
Crossposted-To: comp.unix.programmer
Subject: Re: select on sockets.
Date: Thu, 09 Mar 2000 13:39:58 +0000
tommie jones wrote:
> When I do a select on a set of file descripters (in readfs)
> I not only get a response when a file descriptor is ready to be read
> from but also when a file descriptor is closed by the server (dealing
> with sockets here) what other type of events will select tell me about,
> is there a way to tell which event happened.
I think that you get more information if you use poll().
However if the file has closed I believe that a receive gives zero
bytes of data.
Peter
------------------------------
From: [EMAIL PROTECTED] (Peter Samuelson)
Subject: Re: Creating VB RAD Style Application to Linux
Date: 9 Mar 2000 07:36:32 -0600
Reply-To: Peter Samuelson <[EMAIL PROTECTED]>
[EiNet <[EMAIL PROTECTED]>]
> Where do we want to be with Linux on the next years? In Universities
> for a bunch of guys to enjoy, or in the real world.
In Universities for a bunch of guys to enjoy.
Seriously. I really don't care if Linux achieves Total World
Domination. It does what I want it to, or can be made to if I care to
put in the effort, and basically nothing in my life improves in any
tangible way if Linux makes it really, really big -- except that my
skillset might become a little more valuable.
Why are so many people hung up on "Linux MUST get onto the corporate
desktop" or "Linux MUST beat Windoze 2000" or "Linux MUST put SCO out
of business"? What do any of those things really matter, as long as
there are enough people interested in it to continue to develop cool
stuff with it (which I believe will *never* be a problem) and it
remains freely available?
Peter
------------------------------
From: [EMAIL PROTECTED] (Peter Samuelson)
Subject: Re: Plug and Pray PCI Modem support
Date: 9 Mar 2000 07:40:45 -0600
Reply-To: Peter Samuelson <[EMAIL PROTECTED]>
> > Is there a kernel that supports PCI Plug and Pray modems?
[Alan Donovan <[EMAIL PROTECTED]>]
> First check out whether your modem itself is supported: check out the
> pages below which list all known compatibility issues. Cheap
> "winmodems" delegate a lot of work to the (Windows-only) driver code
> and so won't work at all with Linux as the details are proprietary.
Note also that there are a lot of UARTs (or emulated UARTs) sitting on
PCI busses out there, apparently. Recent kernels support quite a few
varieties. So not all PCI modems are necessarily winmodems. (I
suspect most are, though.)
Peter
------------------------------
From: [EMAIL PROTECTED] (Peter Samuelson)
Subject: Re: Why a file system ?
Date: 9 Mar 2000 07:56:58 -0600
Reply-To: Peter Samuelson <[EMAIL PROTECTED]>
[David Fox <d s f o x @ c o g s c i . u c s d . e d u>]
> My point is that on the balance we would be better off if the order
> of the path elements was not significant.
Assertion only, so far.
> If anyone is interested in going into this topic beyond the level of
> superficial insult, I invite them to suggest a real-world situation
> where a heirarchical file system has advantages over a relational
> file system.
Complexity of implementation? (And lots of peripheral costs associated
with that, such as robustness and efficiency.)
Semi-arbitrary number of "attributes" with no increased complexity of
implementation?
Ease of partitioning namespaces cleanly (for mount points, permissions
management, etc)?
Richer, more expressive namespace? (I.e. you can repeat "attributes"
at different levels if you want.)
Actually I think the burden of proof is on you, who want to change the
status quo. Would a relational filesystem actually buy you anything?
Peter
------------------------------
From: Joe Pfeiffer <[EMAIL PROTECTED]>
Crossposted-To:
alt.comp.hardware,comp.os.linux.hardware,comp.sys.ibm.pc.hardware.misc,comp.sys.ibm.pc.hardware.storage
Subject: Re: Req. for info on rev. engineering an ISA (Backer) card
Date: 09 Mar 2000 08:23:38 -0700
nobody <[EMAIL PROTECTED]> writes:
>
> So... for my next project I think I'll take a crack at reverse
> engineering the interface. What I am looking for is recomendations for
> any software tools that would help me do this. I have a fair bit of
> experience with the ISA interface from both the hardware and software
> end of things so I'm not a novice but I've never hacked someone else's
> work before. The card is currently in my old DOS box and I have several
> little DOS utilties that came with the card that just do some basic I/O
> stuff. For my first attempt I tried watching one of the programs with a
> debugger but that quickly got very tedious.
>
> Any suggestions? If nothing is forthcoming it'll be back to the
> debugger for me.
My answer will be a bit roundabout -- there is a series of books with
names like ``The Undocumented PC,'' ``Undocumented DOS,'' and so on
(they are invaluable for reverse engineering interfaces all by
themselves, incidentally). They have ads in the back for some
programs with names like ``Sourcer.'' Buy those programs -- they are
very nice disassemblers, and are just unbelievably useful for your
task.
I'm the guy who wrote the Compaq Concerto pen driver, so I got a lot
of use out of my copies of these things.
--
Joseph J. Pfeiffer, Jr., Ph.D. Phone -- (505) 646-1605
Department of Computer Science FAX -- (505) 646-1002
New Mexico State University http://www.cs.nmsu.edu/~pfeiffer
VL 2000 Homepage: http://www.cs.orst.edu/~burnett/vl2000/
------------------------------
From: "Mark Tranchant" <[EMAIL PROTECTED]>
Subject: ERROR MESSAGE: Unable to load interpreter
Date: Thu, 9 Mar 2000 15:43:28 -0000
Reply-To: <[EMAIL PROTECTED]>
I have RH6.1 on my 486. I'm not a Linux newbie, having used it since 1.0.9
days. I'm stumped with this one, though:
I've just compiled and installed 2.2.14, replacing 2.2.12-20 that comes with
RH6.1. I use LOADLIN, so I put the newly-created bzImage on my DOS
partition, which is booted from a CONFIG.SYS menu. It boots fine, until init
starts. I get (from memory):
INIT version 2.77 starting (big pause)
Unable to load interpreter (small pause)
NET version... etc.
>From then on the boot is normal and the system appears to work as expected.
Any ideas? I've made as much as possible as modules, which are correctly
installed. ELF and IDE are not modules, however, so it isn't that. I had to
put System.map in /boot to get rid of some error messages, and remove a
modules-info-2.2.12-20 (or something) file in /boot.
Mark.
--
Mark.
------------------------------
From: Charlie Martin <[EMAIL PROTECTED]>
Subject: Re: Recompiled Kernel W/SMP support
Date: Thu, 09 Mar 2000 08:56:12 -0700
Perhaps you have already solved this, but if not: I had the same
problem about 6 months ago, and tried to fix it
via the mkinitrd route. I could never get that to work. But
what finally worked was to configure the scsi driver to
compile into the kernel, rather than having it load as
a module at boot time.
Charlie
Jan Just Keijser wrote:
> In article <8a1eis$h9d$[EMAIL PROTECTED]>, [EMAIL PROTECTED] wrote:
> >I have recompiled the kernel with SMP support enabled (I have a 2 PIII
> >600's), and when I try and boot from the new kernel I get the following
> >error:
> >
> >Loading aic7xxx module
> >/lib/aic7xxx.o: kernel-module version mismatch
> > /lib/aic7xxx.o was compiled for kernel version 2.2.12-20smp
> > while this kernel is version 2.2.12-20.
> >
> >
> >it then panics because it can't find the root file system.
> >
> >
> >
------------------------------
From: Alan Donovan <[EMAIL PROTECTED]>
Subject: Re: memcpy over pci is slow
Date: Thu, 09 Mar 2000 15:50:29 +0000
Charlie Martin wrote:
>
> Within my driver's interrupt service routine, I use a memcpy_fromio()
> to read roughly 64KB from a PCI target device. The memory throughput
> seemed slow, so we looked at the PCI signals on the card using
> an oscope, and found that the transfer rate is something like 2 - 4
> MB/s.
>
> Looking closer, we could see that although the PCI transfers are being
> done in burst mode, the time between bursts is relatively long.
>
> Anyone have suggestions on why the transfer throughput is so slow here?
Is it possible that the local bus (on the far side of your PCI bridge
chip on your card is delivering bytes slowly, and the bridge is caching
up 32 bytes or whatever and then bursting that across the PCI bus at
full speed? Are your local and PCI buses speed matched?
Sorry I can't be more helpful; currently I'm working on a PCI driver
using DMA myself and it's largely new to me.
alan
--
========================================================================
Alan Donovan [EMAIL PROTECTED] http://www.imerge.co.uk
Imerge Ltd. +44 1223 875265
------------------------------
From: Alan Donovan <[EMAIL PROTECTED]>
Subject: Re: Can i get a MAC address ?
Date: Thu, 09 Mar 2000 15:56:27 +0000
[EMAIL PROTECTED] wrote:
>
> In article <38b27e25$0$[EMAIL PROTECTED]>,
> Laurent Gauthier <[EMAIL PROTECTED]> wrote:
> > I'm writing a program that needs to know the MAC address of the
> network
> > device, I think there is a C library that I can use to do it.
> > Can someone help me?
>
> In a shell environment, there is the command "arp"
> You can retrieve a MAC address calling "arp -a <ip-address>"
>
> Within a programm either do some "popen()" or look for an appropriate
> function.
I answered this question a while ago but I can't remember the details of
the correct ioctl. As always, the method is: find a program that does it
(ifconfig in this case), strace it and/or look at the source code, which
can be found automatically by using rpm -qpi and rpmfind.net (if you use
RedHat).
A quick check of man ioctl_list(2) suggests it is:
0x00008915 SIOCGIFADDR struct ifreq * // I-O
You need to pass "eth0" as device name in the ifreq struct.
alan
--
========================================================================
Alan Donovan [EMAIL PROTECTED] http://www.imerge.co.uk
Imerge Ltd. +44 1223 875265
------------------------------
From: [EMAIL PROTECTED] (Stefaan A Eeckels)
Crossposted-To: comp.os.linux.hardware,comp.os.linux.setup
Subject: Re: Impasse with 2 SCSI controllers, kernel mods required?
Date: 9 Mar 2000 16:24:44 GMT
In article <8a7580$[EMAIL PROTECTED]>,
Stephen Lee - Post replies please <[EMAIL PROTECTED]> writes:
> But my guess is still that you have a resource conflict somewhere.
I had exactly the same problem with a 1542 using the same IRQ as
a PCI device (in my case, a network card). The solution was to
reserve IRQ 11 and DMA 5 for ISA legacy devices in the BIOS setup.
--
Stefaan
--
--PGP key available from PGP key servers (http://www.pgp.net/pgpnet/)--
Ninety-Ninety Rule of Project Schedules:
The first ninety percent of the task takes ninety percent of
the time, and the last ten percent takes the other ninety percent.
------------------------------
From: d s f o x @ c o g s c i . u c s d . e d u (David Fox)
Subject: Re: Why a file system ?
Date: 09 Mar 2000 09:24:11 -0800
[EMAIL PROTECTED] (Peter Samuelson) writes:
> [David Fox <d s f o x @ c o g s c i . u c s d . e d u>]
> > My point is that on the balance we would be better off if the order
> > of the path elements was not significant.
>
> Assertion only, so far.
>
> > If anyone is interested in going into this topic beyond the level of
> > superficial insult, I invite them to suggest a real-world situation
> > where a heirarchical file system has advantages over a relational
> > file system.
>
> Complexity of implementation? (And lots of peripheral costs associated
> with that, such as robustness and efficiency.)
>
> Semi-arbitrary number of "attributes" with no increased complexity of
> implementation?
>
> Ease of partitioning namespaces cleanly (for mount points, permissions
> management, etc)?
>
> Richer, more expressive namespace? (I.e. you can repeat "attributes"
> at different levels if you want.)
>
> Actually I think the burden of proof is on you, who want to change the
> status quo. Would a relational filesystem actually buy you anything?
These are good issues, and you may be right. I'm afraid I don't have
the energy to argue about it right now. Consider that a cop-out if
you like.
--
David Fox http://hci.ucsd.edu/dsf xoF divaD
UCSD HCI Lab baL ICH DSCU
------------------------------
From: Marc SCHAEFER <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.linux.development.apps,comp.os.ms-windows.programmer.nt.kernel-mode,comp.unix.programmer,microsoft.public.win32.programmer.kernel
Subject: Re: Tape Driver EOF
Date: 9 Mar 2000 12:55:51 GMT
In comp.os.linux.development.system anonymous <[EMAIL PROTECTED]> wrote:
: However, the problem is that the Unix "dd" cannot find the EOF marker
: and continue to read through the files until physical tape-end.
You don't specify the UNIX version, so I will assume it's Linux 2.2.x
dd will stop (because one of its read() will return 0) on a FILE mark.
A file mark will be created automatically, on UNIX, after a write when
closing the device, or when using the MTWEOF ioctl().
: What is the EOF code that should be wrote to the tape in order for the Unix
: application
: recognize it ? or any other step I have missed ?
maybe your MS-* application uses setmarks, or short records, or
whatever.
------------------------------
From: "Paul D. Smith" <[EMAIL PROTECTED]>
Subject: Re: Swapping HDs... how best to do it?
Date: 09 Mar 2000 13:17:35 -0500
Reply-To: [EMAIL PROTECTED]
%% [EMAIL PROTECTED] (Mike Dowling) writes:
>> It's getting stuffy on there, and really, HD space is _so_ cheap, so I
>> went out and got a new 15G HD.
md> Wow! I cannot get my 2GB disk even half full. My system is around 640
md> MB, just big enough to back up to a CD.
I left 6G of my 15G unpartitioned because I couldn't think of _anything_
do with it... and most of the partitions I _did_ make are empty too
(obviously, since I moved from a 2G disk :).
However, I was definitely out of space at 2G. For my development work I
need Emacs, TeX, the -dev packages for many of the libraries, a full
Perl installation, lots of docs, etc. I had Gnome installed for my wife
who's got a Windows background and can't deal with my tuned FVWM desktop
:).
Etc.
Thanks for all the help from everyone; I installed the new disk last
night and it went perfectly. I ended up using dump/restore from
single-user mode to populate the new disk with the partitions mounted
under "/new", then tweaked the /new/root/etc/fstab, shut down the
machine, took out the old drive and moved the new one to its place,
rebooted, and voila! Worked the first time.
I looked at Norton Ghost, but as far as I could tell the only trial
versions were for Windows, although they said they supported ext2. I
tried GNU parted, but the version in Debian potato just dumped core
immediately in a libc function (I'm upgrading to the latest libc and see
if that fixes it; if not I'll submit a bug).
I decided to use dump/restore instead of tar or cp because I wanted to
un-symlink-ify my filesystem as I moved it over (I was forced to make
some things in /usr symlinks to another filesystem because I ran out of
room on /usr, and I wanted to clean that up) and restore is pretty nice
about letting you extract different sections of the dump in different
places.
So, happiness reigns!
Now, to install a bunch more stuff on my system ;) -- need Samba,
Apache, ... hmm ... maybe Quake ...
--
===============================================================================
Paul D. Smith <[EMAIL PROTECTED]> Network Management Development
"Please remain calm...I may be mad, but I am a professional." --Mad Scientist
===============================================================================
These are my opinions---Nortel Networks takes no responsibility for them.
------------------------------
From: [EMAIL PROTECTED] (M Sweger)
Subject: How to set kernel name? I have (none).
Date: 9 Mar 2000 19:20:12 GMT
Reply-To: [EMAIL PROTECTED]
Hi,
How does one set the kernel name? In my syslog output the
kernel name is (none) kernel: some message.
My uname -s is Linux. I've even set the profile so that BOOT_IMAGE
is set to Linux in the environment like the man page for lilo says. Note:
I don't have lilo or ext2, I have UMSDOS.
Thanks
--
Mike,
[EMAIL PROTECTED]
------------------------------
From: "Raj Suri" <[EMAIL PROTECTED]>
Subject: kmalloc allocate swappable memory?
Date: Thu, 9 Mar 2000 13:52:49 -0700
This is a multi-part message in MIME format.
=======_NextPart_000_0009_01BF89CE.C1F30220
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
I was curious as to whether or not the Linux kernal can swap pages of =
memory to disk which were allocated using kmalloc in kernel space. Or =
does that memory have to stay in physical memory? Is there any resource =
that explains this in detail??
=======_NextPart_000_0009_01BF89CE.C1F30220
Content-Type: text/html;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META content=3D"text/html; charset=3Diso-8859-1" =
http-equiv=3DContent-Type>
<META content=3D"MSHTML 5.00.2919.6307" name=3DGENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=3D#ffffff>
<DIV><FONT size=3D2>I was curious as to whether or not the Linux kernal =
can swap=20
pages of memory to disk which were allocated using kmalloc in kernel=20
space. Or does that memory have to stay in physical memory? =
Is there=20
any resource that explains this in detail??</FONT></DIV></BODY></HTML>
=======_NextPart_000_0009_01BF89CE.C1F30220==
------------------------------
From: Surya <[EMAIL PROTECTED]>
Subject: device or resource busy ERROR
Date: Wed, 08 Mar 2000 02:12:16 -0800
Reply-To: [EMAIL PROTECTED]
Hi all,
I wrote a module and when I try to install it, it gives an error
message "Device or Resource busy". Is it because the initi_module failed
to intialize? How to trace the cause of the error. In Init_module I
allocate memory in the kernel using kmalloc() and use a printk to report
error if kmalloc() fails. But I don't see any printk messages in
/var/log/messages file. So what could be the problem. Any suggestions or
comments are welcome.
Thanks in anticipation,
Surya.
------------------------------
From: Surya <[EMAIL PROTECTED]>
Subject: init_module: device or resource busy error
Date: Thu, 09 Mar 2000 04:08:23 -0800
Reply-To: [EMAIL PROTECTED]
Hi,
I am getting this error when I insmod my module. My module allocates
memory for a struct using kmalloc( xxx, yyyy, GFP_KERNEL) in the
init_module() function. In init_module() I track memory allocation using
printk. I find that the memory has been allocated. Also I create a proc
file entry, which is also being created. But the module is not
installed, it returns with an error "init_module: device or resource
busy" message. What is the exact meaning of this error message. Can
someone explain the reason and how to avoid it.......
Thanx in anticipation,
Surya.
------------------------------
From: Surya <[EMAIL PROTECTED]>
Subject: Re: init_module: device or resource busy error
Date: Thu, 09 Mar 2000 06:19:59 -0800
Reply-To: [EMAIL PROTECTED]
Everything is OK now................
I made a simple mistake i.e. I returned 1 instead of 0 for success and
hence the init_module() is showing that error message.
The correct code should look like....
int init_module()
{
blah..... blah..... //device driver initialization stuff
if(success)
return 0;
else{
////// remove any previously initialized stuff here otherwise they
will remain forever
return 1;
}
}
Then everything should be OK.
Surya.
------------------------------
** 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
******************************
