Linux-Development-Sys Digest #428, Volume #6 Thu, 25 Feb 99 08:14:00 EST
Contents:
Re: ncurses4.2/egcs1.1.1/glibc2.1 compile problem ("T.E.Dickey")
Tape still not working with ncr53c8xx driver? (bill davidsen)
Re: Another process control - type question
(=?iso-8859-1?Q?Jos=E9=20Mar=EDa=20Fern=E1ndez=20Gonz=E1lez?=)
SMC Ether EZ driver... (Robspiere)
Re: net-pf- [was Re: net-pf-17 ?] (Kip Rugger)
Re: Will 2.2.x support removable medias better? (Peter Samuelson)
Re: Why I'm dumping Linux, going back to Windblows (Sniper)
Re: threads (Michael Hirsch)
Re: compile problem bind-8.1.2 for Linux 2.2.1 & glibc2.1 (Remco van den Berg)
Re: Why I'm dumping Linux, going back to Windblows (Sniper)
Nr of loopdevices (Thomas Rink)
Apache and persistent CGI programs????? ("David Sisk")
Re: Problems compiling linux-2.2.2 (Usseglio Gaudi Francesco)
Re: Overclocking (was: Re: K6-2 and Linux, Are there any Bug?) (Yair Paz)
Re: Overclocking (was: Re: K6-2 and Linux, Are there any Bug?) (Michael Creasy)
Re: IP to process network interface? (Frank Sweetser)
----------------------------------------------------------------------------
From: "T.E.Dickey" <[EMAIL PROTECTED]>
Subject: Re: ncurses4.2/egcs1.1.1/glibc2.1 compile problem
Date: 25 Feb 1999 01:11:22 GMT
Dirk Lattermann <[EMAIL PROTECTED]> wrote:
> Hello,
> I don't really know where to post this, but this might be a
> right place:
> has someone successfully built ncurses-4.2 with glibc2.1
> and egcs-1.1.1? Are there known problems?
your problem really looks like the awk/sed programs are getting broken
by confusion between old/new shared libraries. (I don't know of anything
offhand that would cause the script to drop characters).
> #undef pe
> ( pe z)
> {
> lx)"), (long)z)); returnCode( pe(z));
> }
--
Thomas E. Dickey
[EMAIL PROTECTED]
http://www.clark.net/pub/dickey
------------------------------
From: [EMAIL PROTECTED] (bill davidsen)
Subject: Tape still not working with ncr53c8xx driver?
Date: 24 Feb 1999 23:16:31 GMT
Running 2.2.1 release, three systems, all have old or new DAT tapes
(SCSI-1 or 2), one has an old drive for DC150 tapes to read old backups,
one has a 5GB Exabyte with a type I can't readily read.
What they have in common is that none of the tapes work with the
bcr53c8xx driver, and all work with the 53c7,8xx driver. That driver is
slower for disk by a bit, I'd rather not use it.
Oh, one P5-133S, one P5-200MMX, one 6X86-300.
I noted this back on 2.1.131, and several people posted "me too"
messages. Is someone looking at this? Is there a trick?
--
bill davidsen <[EMAIL PROTECTED]> CTO, TMR Associates, Inc
Politicians and diapers have one thing in common. They should both be
changed regularly and for the same reason.
--Ted Symons(?)
------------------------------
From: =?iso-8859-1?Q?Jos=E9=20Mar=EDa=20Fern=E1ndez=20Gonz=E1lez?=
Subject: Re: Another process control - type question
Date: Wed, 24 Feb 1999 20:36:06 +0100
David Sisk wrote:
>
> OK, I'm going to draw on some programming that I did years ago on an IBM
> AS/400 using RPG (I know, I know, EEEEEE!!!) to pose a particular question:
>
> It was possible to have program1 call program2 with parms like so (in
> pseudo-code):
>
> * program1 source...
> ...
> call "program2 somedata n" -- Load program2, do some work, and return the
> parms
> call "program2 somemoredata n" -- program2 already loaded, do some more
> work, return the parms
> call "program2 nodate y" -- tell program2 to exit.
> ..
>
> * program 2 source
> get the parms
> do some work
> If parm2 = 'n' RETURN -- don't exit, stay loaded
> else SETON *LR -- exit and unload the program (meaning about the same thing
> as EXIT(0) in C).
>
> This was a very efficient way to call an external program because it stayed
> loaded until you sent it the parms to tell it to exit.
>
> Question:
>
> Is it possible to do the same sort of thing (using C) in Linux (or any Unix
> for that matter)? How about in Win32?
> If the exact same type of method is not available, what's a good way to
> emulate it?
>
> Any responses appreciated. Doubly thankful if you'd email as well as
> posting.
>
> Regards,
>
> --
> David C. Sisk
> The Unofficial ORACLE on NT site
> http://www.ipass.net/~davesisk/oont.htm
Why don't you play with IPC and other related stuff (semaphores, shared memory,
etc...) ?
On Win32, you could detect if there is another running program instance through
the CreateMutex call...
--
José María
------------------------------
From: [EMAIL PROTECTED] (Robspiere)
Subject: SMC Ether EZ driver...
Date: 25 Feb 1999 05:51:58 GMT
I installed Red Hat 5.2 (kernel 2.0) about two weeks ago, and I've been looking
ever since for a driver for my SMC ethernet card. The SMC web site has a
utility that will supposedly set the whole thing up, but when I run it, it
returns the oh-so-helpful "errno22". The documentation says nothing of error
numbers, BTW.
The card is the SMC Ether EZ, model 8614T. It's ISA, and PnP. I've got a
Micron Pentium 166.
Any help you could provide would be greatly appreciated.
Thanks in advance. I would appreciate it if you could CC any replies to me via
email, in addition to posting them to the group.
Rob
------------------------------
From: [EMAIL PROTECTED] (Kip Rugger)
Subject: Re: net-pf- [was Re: net-pf-17 ?]
Date: 12 Feb 1999 17:10:46 -0600
>I don't think that it is modprobe that is deciding to
>look for net-pf-; I've looked in the source for the sys_socket
>call and it has a sprintf(buf,"net-pf-%d",....) line. So
- ifconfig does socket(AF_IPX, ...) which expands to socket(4, ...)
- kernel cannot find protocol 4, so launches "modprobe net-pf-4"
- modprobe looks in /etc/conf.modules to get module name associated with
net-pf-4
case 1: modprobe finds "alias net-pf-4 ipx"
modprobe will load the ipx stack and exit
Or will fail to find ipx and print the annoying
modprobe: can't locate module net-pf-4
case 2: modprobe finds "alias net-pf-4 off"
modprobe does nothing
case 3: modprobe does not find "alias net-pf-4 ..."
modprobe uses a builtin default, "alias net-pf-4 ipx"
and proceeds as in case 1
- kernel checks again to see if protocol 4 is present, and either
calls the module to handle it or returns an error to the original
socket(4, ...) call.
>Where is this file that you mention wherein the aliases are kept ?
/etc/conf.modules
It is optional, so create it and add aliases as required.
See man modprobe.
------------------------------
From: [EMAIL PROTECTED] (Peter Samuelson)
Subject: Re: Will 2.2.x support removable medias better?
Date: 25 Feb 1999 01:31:42 -0600
Reply-To: Peter Samuelson <[EMAIL PROTECTED]>
> > > The point is that it should be unnecessary to mount/umount a
> > > removable media if the OS is capable of detecting
> > > insertion/ejection of media.
[Horst von Brand <[EMAIL PROTECTED]>]
> > Can't be done in general, due to the extremely clever PC
> > architecture.
[Adam P. Jenkins <[EMAIL PROTECTED]>]
> Can you please elaborate? I don't remember ever having to mount or
> unmount a CDROM, Zip, Jaz, or floppy disk under Windows.
I think you misunderstood the phrase "can't be done in general". This
was in response to "if the OS is capable of detecting
insertion/ejection of media". The point is that most removable media
drives for PC's do not provide any way to notify the OS that the media
has been changed. In some cases I think you can poll for that
information, one way or another -- Windoze seems to for CD's. For a
standard floppy drive I don't think you can get a straight answer
without actually spinning up the drive motor.
Certainly it is possible to use removable media without explicit mount
and unmount commands -- see Stephen Tweedie's supermount patch for
2.0.x (which I wish someone would port to 2.2 one of these days but
don't care quite enough to do it myself). But I think supermount
detects a media change by detecting *content* change. Not the same
thing as if, for example, your floppy drive could just post an
interrupt whenever you opened or closed the drive door.
--
Peter Samuelson
<sampo.creighton.edu!psamuels>
------------------------------
From: [EMAIL PROTECTED] (Sniper)
Crossposted-To: alt.os.linux,comp.os.linux.development.apps,comp.os.linux.setup
Subject: Re: Why I'm dumping Linux, going back to Windblows
Date: Tue, 23 Feb 1999 14:50:45 GMT
Reply-To: [EMAIL PROTECTED]
On Tue, 23 Feb 1999 03:43:23 +0100, Emile van Bergen
<[EMAIL PROTECTED]> wrote:
>jedi wrote:
>
>> >As for sitting running on a 486 with 24 megs of memory as a
>> >firewall/NNTP/SMTP server, then yes ! But NT can do that as well, may
>> >not quite as well, but definately easier to set up out of the box.
snip
>
>Oh, and my desktop has a mere 32Mb to run X, Netscape, Staroffice (which
>imports Word '97 documents correctly, has a HTML WYSIWYG editor,
>spreadsheet, Planner; the personal edition is free, BTW), and does quite
>some gcc work.
>
>Vmstat on both machines don't ever show 'swapped pages per second' above
>zero for longer than a second (and it occurs rarely), so I notice no
>sluggishness due to swapping.
>
>I don't even _want_ to say, try that with NT. I think I don't need to.
Can you export back to 97 format as well ? What about Excel 97 ?
"What's the difference between a nail, a screw and a bolt ?" the woodwork
teacher asked the only girl in the class during the first day of school.
She pondered the question for a moment, then replied, "Well, I can't rightly
say as I know, 'cause I ain't never been 'bolted'."
------------------------------
From: Michael Hirsch <[EMAIL PROTECTED]>
Subject: Re: threads
Date: 23 Feb 1999 09:23:17 -0500
Jim Cromie <[EMAIL PROTECTED]> writes:
> Linux threads are provided in Kernel space, as opposed to user space. Though
> I dont know whether that means you can actually create new threads in the
> kernel.
Isn't that what bdflush is (and kmod too, perhaps)?
> Actually, no, system calls are still atomic, and cant be pre-empted by another
> call,
> therefore theres no use for extra kernel threads. the Threads-FAQ (or other
> stuff
> Ive read tonite) indicates that re-entrant system calls are in progress in the
> 2.3
> (development kernel)
I thought that with version 2.2 there are lots of independent system
calls. So even though one system call can't run twice simultaneously,
two different system calls could if they are dealing with very
different parts of the kernel.
--
Michael D. Hirsch Work: (404) 727-7940
Emory University, Atlanta, GA 30322 FAX: (404) 727-5611
email: [EMAIL PROTECTED] http://www.mathcs.emory.edu/~hirsch/
Public key for encrypted mail available upon request (or finger
[EMAIL PROTECTED]).
------------------------------
From: Remco van den Berg <[EMAIL PROTECTED]>
Subject: Re: compile problem bind-8.1.2 for Linux 2.2.1 & glibc2.1
Date: Tue, 23 Feb 1999 15:56:34 +0100
Andreas Jaeger wrote:
>
> >>>>> Remco van den Berg writes:
>
> > Is there somebody around having a patch for bind-8.1.2 ?
>
> > I'm having problems compiling it under Linux 2.2.1 and glibc2.1.
>
> > make[2]: Entering directory `/usr/src/Network/bind-8.1.2/lib/isc'
> > gcc -O -g -I../../port/linux/include -I../../include -c eventlib.c
> > eventlib.c: In function `__evGetNext':
> > eventlib.c:293: structure has no member named `fds_bits'
> > eventlib.c:294: structure has no member named `fds_bits'
> > eventlib.c:295: structure has no member named `fds_bits'
>
> > Problem seems to be that the structure fd_set doesn't have a
> > member fds_bits. Even #defining __USE_XOPEN doesn't help.
> You can't define __USE_XOPEN. Check <features.h> for the defines you
> can add and try _XOPEN_SOURCE.
>
> Andreas
I downloaded the beta release of bind (bind-8.2) and changed the
file port_before.h to:
%> cat port/linux/include/port_before.h
#undef WANT_IRS_NIS
#undef WANT_IRS_PW
#undef WANT_IRS_GR
#define SIG_FN void
/* Added for GLIBC2.1 support: */
#define HAS_INET6_STRUCTS
#define _GNU_SOURCE
#include <features.h>
/* end addition */
#if defined(HAS_PTHREADS) && defined(_REENTRANT)
#define DO_PTHREADS
#endif
/* We grab paths.h now to avoid troubles with redefining _PATH_HEQUIV later.
*/
#include <paths.h>
And now everything compiled fine! (I hope it will run fine also :-)
Thanks for the help,
-Remco van den Berg
------------------------------
From: [EMAIL PROTECTED] (Sniper)
Crossposted-To: alt.os.linux,comp.os.linux.development.apps,comp.os.linux.setup
Subject: Re: Why I'm dumping Linux, going back to Windblows
Date: Thu, 25 Feb 1999 10:34:01 GMT
Reply-To: [EMAIL PROTECTED]
On 24 Feb 1999 15:10:08 -0800, d s f o x @ c o g s c i . u c s d . e d
u (David Fox) wrote:
Hit the nail straight on the Head my friend !
Some of these people may also be paying you for it as well.
I may love writing Tex documents (When I go over to linux) but if I
give that to my clients, they are just going to go "Eh ? Whats did den
?" (C)
:->
>Julian Robert Yon <[EMAIL PROTECTED]> writes:
>
>> Matthew Isleb wrote:
>>
>> > Can I ask what this obsession with the Office 97 file format is? I mean,
>> > so you have to save files in an older format for portability, big deal.
>> > Or do the hypothetical users you are talking about use 100% of the Office 97
>> > features which would prevent the usage of another file format. Get real.
>>
>> And of course, Office is the only "Productivity Suite" I have used which
>> can make a simple spreadsheet, a couple of charts and a 300 word report
>> take up 3.5M of disk space. I think our friend is clutching at straws...
>
>My understanding is that in the real world people will send you files
>in Office 97 format which you are expected to be able to read. I know
>it sounds rude, but its true.
.
"What's the difference between a nail, a screw and a bolt ?" the woodwork
teacher asked the only girl in the class during the first day of school.
She pondered the question for a moment, then replied, "Well, I can't rightly
say as I know, 'cause I ain't never been 'bolted'."
------------------------------
From: Thomas Rink <[EMAIL PROTECTED]>
Subject: Nr of loopdevices
Date: Tue, 23 Feb 1999 17:00:59 +0100
Reply-To: [EMAIL PROTECTED]
Hi,
we need more than 8 loop block devices (Kernel 2.0.36).
In /usr/src/linux/driver/block.loop.c there is a macro MAX_LOOP which
seems to hardcode the number of loop devices. Can this number be
increased to e.g. 16 or are there side effects to consider?
TIA, Thomas
------------------------------
From: "David Sisk" <[EMAIL PROTECTED]>
Subject: Apache and persistent CGI programs?????
Date: Tue, 23 Feb 1999 18:58:39 -0500
Hi:
Another quick question (I hope!). Is it possible for Apache (or any
CGI-compliant web-server, for that matter) to launch a process in the normal
CGI manner but NOT kill it until later? I understand that the CGI program
would have to be written to not end immediately, but to sit and wait for
further instructions (maybe via named pipes or something along those lines).
If the process didn't die, is there a CGI-based method to send it further
messages.
For example:
Hit1: launch the CGI program
Hit2: do operation1
Hit3: do operation2
Hit4: exit the CGI program.
Any ideas on how to point the instructions back to the right existing
process?
Is this possible with the current CGI specification? Any comments truly
appreciated. Doubly thankful if you'd email as well as posting.
Regards,
--
David C. Sisk
The Unofficial ORACLE on NT site
http://www.ipass.net/~davesisk/oont.htm
------------------------------
From: Usseglio Gaudi Francesco <[EMAIL PROTECTED]>
Subject: Re: Problems compiling linux-2.2.2
Date: Thu, 25 Feb 1999 12:09:55 +0100
Friedhelm Hinrichs wrote:
>
> Compiling the new kernel 2.2.2 ended with the following messages (I hope
> this does make sense to somebody):
>
> gcc -D__KERNEL__ -I/usr/src/linux/include -Wall -Wstrict-prototypes -O2
> -fomit-frame-pointer -pipe -fno-strength-reduce -m486 -malign-loops=2
> -malign-jumps=2 -malign-functions=2 -DCPU=586 -c -o loopback.o
> loopback.c
> /usr/src/linux/include/net/sock.h: In function `sk_filter':
> In file included from loopback.c:51:
> /usr/src/linux/include/net/sock.h:796: dereferencing pointer to
> incomplete type
> /usr/src/linux/include/net/sock.h:796: dereferencing pointer to
> incomplete type
> /usr/src/linux/include/net/sock.h:796: warning: passing arg 1 of
> `sk_run_filter' from incompatible pointer type
> /usr/src/linux/include/net/sock.h:796: too few arguments to function
> `sk_run_filter'
> /usr/src/linux/include/net/sock.h: In function `sk_filter_release':
> /usr/src/linux/include/net/sock.h:807: warning: implicit declaration of
> function `sk_filter_len'
> /usr/src/linux/include/net/sock.h:811: dereferencing pointer to
> incomplete type
> /usr/src/linux/include/net/sock.h: In function `sk_filter_charge':
> /usr/src/linux/include/net/sock.h:817: dereferencing pointer to
> incomplete type
> make[3]: *** [loopback.o] Error 1
> make[3]: Leaving directory `/usr/src/linux-2.2/drivers/net'
> make[2]: *** [first_rule] Error 2
> make[2]: Leaving directory `/usr/src/linux-2.2/drivers/net'
> make[1]: *** [_subdir_net] Error 2
> make[1]: Leaving directory `/usr/src/linux-2.2/drivers'
> make: *** [_dir_drivers] Error 2
>
> --
Hi,
I had the same problem. I resolved it unchecking the "IP filtering" (or
something similar:socket filter) and the problem disappeared. I think it's a
bug in the 2.2.2 because the 2.2.1 works perfectly with that option on: i dig a
little in the kernel source and i find that the function prototype of
sk_run_filter has been changed. (btw i tried to patch it by i wasn't able to
find where the sk_filter struct is defined).
bye.
--
==+==
...era un mondo adulto...
...si sbagliava da professionisti.
Paolo Conte
------------------------------
From: Yair Paz <[EMAIL PROTECTED]>
Crossposted-To:
alt.os.linux.slackware,comp.os.linux.hardware,comp.os.linux.misc,comp.os.linux.setup
Subject: Re: Overclocking (was: Re: K6-2 and Linux, Are there any Bug?)
Date: Thu, 25 Feb 1999 11:46:25 +0200
Reply-To: [EMAIL PROTECTED]
Bugs Bunny wrote:
> you can overclock it it will run hotter and finally burn out like all chips
> that are over clocked it is a gamble
>
> do it aown risk
Bulls. some chips are meant to beoverclocked - the P2-300 for example,
in the last 6 months of it's manufacture, it was made with a P2-350 core
hard wired to be 300. you can easily un-wire it (OK , not so easily). or
the Celeron 300A which overclocks to 450 w/o a glitch just by changing
the FSB speed, or the K6-2 300 which can overclock to 375 w/o additional
cooling, or any P5 MMX chips which can usually be overclocked to about
20% more, if you have the right board, w/o needing for extra hardware
(coolants) or voltage tweeking. nowdays, you can even overclock
HardDisks.
Oded
------------------------------
From: Michael Creasy <[EMAIL PROTECTED]>
Crossposted-To:
alt.os.linux.slackware,comp.os.linux.hardware,comp.os.linux.misc,comp.os.linux.setup
Subject: Re: Overclocking (was: Re: K6-2 and Linux, Are there any Bug?)
Date: Thu, 25 Feb 1999 12:26:51 +0000
If this is the case can someone tell me why my K6-2 350 when overclocked
to 400 crashes linux on boot, it has a huge fan on it.
Michael
> Bulls. some chips are meant to beoverclocked - the P2-300 for example,
> in the last 6 months of it's manufacture, it was made with a P2-350 core
> hard wired to be 300. you can easily un-wire it (OK , not so easily). or
> the Celeron 300A which overclocks to 450 w/o a glitch just by changing
> the FSB speed, or the K6-2 300 which can overclock to 375 w/o additional
> cooling, or any P5 MMX chips which can usually be overclocked to about
> 20% more, if you have the right board, w/o needing for extra hardware
> (coolants) or voltage tweeking. nowdays, you can even overclock
> HardDisks.
> Oded
------------------------------
From: Frank Sweetser <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.networking
Subject: Re: IP to process network interface?
Date: 24 Feb 1999 16:01:10 -0500
[EMAIL PROTECTED] (Phil Howard) writes:
> Is there a way to get an IP network interface into the kernel so
> that the kernel can operate on it as any network interface, but
> have the interface really connect via a process on that machine?
check out the 2.2 kernels, i vaguely recall some features along these lines
being introduced.
--
Frank Sweetser rasmusin at wpi.edu fsweetser at blee.net | PGP key available
paramount.ind.wpi.edu RedHat 5.2 kernel 2.2.1 i586 | at public servers
You want it in one line? Does it have to fit in 80 columns? :-)
-- Larry Wall in <[EMAIL PROTECTED]>
------------------------------
** 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
******************************
