Linux-Development-Sys Digest #399, Volume #6 Sat, 13 Feb 99 18:14:44 EST
Contents:
Re: glibc-2.1 compiled but crashed my System (Thorsten Kukuk)
Re: glibc 2.1 ;) (Thorsten Kukuk)
Re: glibc 2.1 ;) (Peter Mardahl)
Linux on old as400 machines?
Re: Help : Time Measurement (Bengt Richter)
@@GLIBC (Dan Shechter)
Re: glibc 2.1 ;) (Jack Howarth)
Re: K6-400 "kernel paging request" errors (Jose Urena)
Can I extend struct kernel_stat? (Kalle Olavi Niemitalo)
Re: Why I'm dumping Linux, going back to Windblows (Garson Cole)
Re: Problem with autofs and local /home (Craig J Copi)
graphics lib? (Ricky Nelson)
__setfpucw disappeared from glibc (Michal Szymanski)
Re: @@GLIBC (Juergen Heinzl)
----------------------------------------------------------------------------
From: Thorsten Kukuk <[EMAIL PROTECTED]>
Subject: Re: glibc-2.1 compiled but crashed my System
Date: 13 Feb 1999 09:58:33 GMT
[EMAIL PROTECTED] wrote:
> Hi!
> I just compile the new glibc-2.1 with
> ./glibc-2.1/configure --prefix/usr --e
This wasn't the configure command you run, or ?
It should be configure --prefix=/usr --enable-add-ons=yes
> make
> make check
> all without errors. :)
> But make install fails on the final perl-script, which should check the install..
> And from now most apps fails with the same error and the system won't start anymore.
> hmm, don't know the exact output at the moment, but it was like ...
> _dl_global_scope has different sizes in hared libryries
> _dl_default_scope ->undefined symbol in shared libraries
It seems you have a mixed set of /lib/ld-linux.so.2 from glibc 2.0.x and
libc.so.6 from glibc 2.1 or ld-linux.so.2 from glibc 2.1 and libc.so.6
from 2.0.x.
Please double check if the ld-linux.so.2 link the libc.so.6 link
shows to the right version.
Thorsten
--
Thorsten Kukuk http://home.pages.de/~kukuk/ [EMAIL PROTECTED]
SuSE GmbH Schanzaeckerstr. 10 90443 Nuernberg
Linux is like a Vorlon. It is incredibly powerful, gives terse,
cryptic answers and has a lot of things going on in the background.
------------------------------
From: Thorsten Kukuk <[EMAIL PROTECTED]>
Subject: Re: glibc 2.1 ;)
Date: 13 Feb 1999 10:03:30 GMT
Nathan Paul Simons <[EMAIL PROTECTED]> wrote:
> 1) Netscape - kinda weird. Netscape won't connect to any
> external sites anymore, but works with loopback.
Which netscape version do you use ? netscape communicator 4.5 runs fine
for me.
> 3) RPM - How's this for a way to screw yerself over: make a
> src rpm of a library, compile the src rpm, then install/upgrade
> it. Find out you want to remove it because it breaks a few things,
> but you can't remove it because rpm, even though it is a static
> executable, depends on certain libraries - the ones you just
> replaced. My backup of rpm-2.5.5.tar.gz is compiling as I
> type . . .
> The exact error was something like "Bad owner/group" and the only
> reason I knew it was the library (besides the fact that I just
> upgraded it) was that I did an strace to find out that certain
> library files it was looking for weren't missing exactly, just
> the version numbers had changed
> (/lib/libnss_compat.so.1 -> /lib/libnss.so.2).
Please try my nss-v1 add-on from http://home.pages.de/~kukuk/linux/glibc.html
But with this add-on, the perl testscript from make install will fail.
> Needles to say, I'm not waiting out to find out what else it broke,
> and I wouldn't recommend that anyone else try this anytime soon, especially
> because of the licensing problems. Ah, it's done compiling. Time to
> 'rpm -U --force glibc-2.0.7.i386.rpm'.
>From the announcement:
*BUT*: updating the C library is no trivial task and it is very easy
to damage one's system. Therefore, persons who do not exactly know
what to do, should consider using a binary distribution instead, when
they become available.
Maybe you should wait for a binary distribution ?
Thorsten
--
Thorsten Kukuk http://home.pages.de/~kukuk/ [EMAIL PROTECTED]
SuSE GmbH Schanzaeckerstr. 10 90443 Nuernberg
Linux is like a Vorlon. It is incredibly powerful, gives terse,
cryptic answers and has a lot of things going on in the background.
------------------------------
From: [EMAIL PROTECTED] (Peter Mardahl)
Subject: Re: glibc 2.1 ;)
Date: 12 Feb 1999 22:55:12 -0800
In article <[EMAIL PROTECTED]>,
Paul D. Smith <[EMAIL PROTECTED]> wrote:
>%% [EMAIL PROTECTED] (Nathan Paul Simons) writes:
>
> nps> Now, in case you haven't heard, glibc 2.1 has been pulled due to
> nps> licensing problems and won't be available for a while.
>
>This is completely wrong. I don't know where this rumor got started,
>but the reason glibc 2.1 was pulled from ftp.gnu.org has absolutely
>nothing to do with any licensing.
>
>The real reason is even... stranger...
>
I'd love to know the real reason. I saw the "licensing rumor" on
www.slashdot.org.
PeterM
------------------------------
From: <[EMAIL PROTECTED]>
Subject: Linux on old as400 machines?
Date: Thu, 11 Feb 1999 19:13:05 +0100
=====BEGIN PGP SIGNED MESSAGE=====
Hi,
I'm looking if there is any option to reuse old as400 machines with any
free Un*x OS.
Any hint/link/suggestion/etc.. will be greatly appreciated
Thanks in advance,
Ulisses
PD: please CC your reply to [EMAIL PROTECTED]
- -----------------------------------------------------------------------------
"Computers are useless. They can only give answers." Pablo Picasso
=====BEGIN PGP SIGNATURE=====
Version: 2.6.3a
Charset: latin1
Comment: PGP public key available at http://www.rediris.es/cert/keyserver
iQB1AwUBNsMdtQ/N+5+NQ63pAQH/1AMAuVvZaWSJ9lVk+PEbbWsCahaPgHG2cnf0
1MqgTe7trnzMG9FwXX3Ja5B+5+5ISQMfTquJt48Y4rn4wiR9GtiqqpfqGSMrQDtC
mJ5rk+Z/g7C6oegwstuFG6wp48bXtElx
=pr5c
=====END PGP SIGNATURE=====
------------------------------
From: [EMAIL PROTECTED] (Bengt Richter)
Subject: Re: Help : Time Measurement
Date: Sat, 13 Feb 1999 12:51:54 GMT
On 12 Feb 99 10:15:07 +0100 (MET), [EMAIL PROTECTED]
(Ja~nez Fernandez, Ruben) wrote:
>Hi all:
>
>First of all, excuse my english, just a little too bad :-).
>
>I`m making my final career project under redhat linux 4.2.
>One of the first steps I have to do is the measurement of the real time
>interval between two consecutive executions of one function in a "for" loop.
>My problem is the multitask execution. Does anyone know if there
>is a way to accurately measure that time?
>
>Thanks in advance... :-)
>
> Beto
I'm assuming the "multitask execution" problem is that your
loop is getting interrupted for other processing. So stop playing
with the mouse ;-) Just kidding.
If you have a Pentium of the right model (not too early), your might
try writing a machine language function to read the time stamp
register, which counts CPU clock ticks as a 64-bit integer.
This gives you super accuracy, but you want to time your function
many times. Call the time stamp function before and after the call
to your function, and store your times in an array (so you don't do
a lot of processing that will disturb the chip cache). Then compute
and print the intervals (or sort and graph them if you want to get
fancy). You will find a shortest time for when the cache
is preloaded (i.e., not the first time), a little longer time for the
first time, and some really long ones when your loop gets interrupted.
(I am assuming you are looping fast enough to complete two loops
between interrupts at least some of the time). Throw out the long
ones, unless you want an average under whatever conditions you
are running -- but the long ones will depend on mouse interrupts,
background disk stuff, modem stuff, real time clock ticks (glacial
compared to CPU ticks), etc. and whether and what other processes
and threads get switched to when the interrupts happen,
so you'll have to judge how much and what the averages mean
if you average everything..
There will of course be an overhead in calling the time stamp
function itself (mine was 23 ticks), so you can put two calls in
a loop with nothing between to see what you get, then subtract
that from your function timings.
If your function is complex, you may get variations according to
different logic branches etc., but you will even measure differences
in such things as multiply by zero vs something non-zero, since
instructions themselves sometimes have variable timing. The time stamp
instruction is RDTSC (op code 0F31, which puts the 64-bit
counter in EDX:EAX). I wouldn't think linux would disable it from
access from user mode (TSD bit in CR4), but you'll have to see.
Good luck.
Regards,
Bengt Richter
BTW, I haven't got my Linux going yet, so I can't give you a linux
function for RDTSC, but someone has probably done this.
BTW2, if you don't have a pentium, you can still get pretty accurate
timings by looping thousands of times and counting the number of
times you get particular time values (from a lower resolution timeer)
and interpreting the counts as probabilities of sampling time within
respective clock tick intervals. This works best if your loop start
time is _not_ somehow synchronized with clock edges. Anyway,
you can usually figure how big a fractional tick overlap your interval
has by noting the relative probability of two adjacent timing values.
------------------------------
Date: Sat, 13 Feb 1999 16:29:32 +0200
From: Dan Shechter <[EMAIL PROTECTED]>
Subject: @@GLIBC
Hi,
I have a Redhat 5.1 system + Glibc system
Recently I've noticed that all kinds of RPM's aren't functioning
properly because they fail to find symbols from libc.so, which
have a "wierd" @@GLICC2.0 suffix...
Can someone explain what is the meaning of this?
The current version of GLIBC I have, has no @@GLIBC2.0, should I get
a newer one?, If so, which? Are the older symbols still supported, e.g.,
If a shared object or a binary uses bzero instead of bzero@@GLIBC2.0,
will it still find it with newer glibc's?, Why was this done? How?
Shechter.
------------------------------
From: [EMAIL PROTECTED] (Jack Howarth)
Subject: Re: glibc 2.1 ;)
Date: Sat, 13 Feb 1999 12:34:51 -0500
The political reasons are that egcs has come to be the preferred
development system and is totally outpacing gcc in development and
stability. During the glibc 2.1 development cycle I witnessed a
constant exchange of information between the glibc and egcs developers.
Whereas the gcc folks sit on their fannies and expect to be
"taken care of". Now glibc 2.1 is out with egcs as the recommended
development tool and it appears to stick in Stallman's craw. Well
too bad. This crap where the GNU leadership leverages the efforts
of the GNU development community and then like Big Brother tells
them that gcc is better for them has to stop.
Jack
In article <7a37kg$ie9$[EMAIL PROTECTED]>,
[EMAIL PROTECTED] (Peter Mardahl) wrote:
>In article <[EMAIL PROTECTED]>,
>Paul D. Smith <[EMAIL PROTECTED]> wrote:
>>%% [EMAIL PROTECTED] (Nathan Paul Simons) writes:
>>
>> nps> Now, in case you haven't heard, glibc 2.1 has been pulled
due to
>> nps> licensing problems and won't be available for a while.
>>
>>This is completely wrong. I don't know where this rumor got started,
>>but the reason glibc 2.1 was pulled from ftp.gnu.org has absolutely
>>nothing to do with any licensing.
>>
>>The real reason is even... stranger...
>>
>
>I'd love to know the real reason. I saw the "licensing rumor" on
>www.slashdot.org.
>
>PeterM
------------------------------
From: Jose Urena <[EMAIL PROTECTED]>
Crossposted-To: linux.dev.kernel,comp.os.linux.hardware,comp.os.linux.misc
Subject: Re: K6-400 "kernel paging request" errors
Date: Fri, 12 Feb 1999 16:03:34 -0500
I have done that to keep a few pentiums from stoping, and even my K6-200
It could be the CPU not cooling fast enough or something else on the board that it's
getting
offset by the heat released by the CPU
Christian Aasland wrote:
> Another thought - my roomate had a problem with his machine locking up after a few
> minutes of q2, figured out it was a heat problem. He solved the problem by taking
> the cover off his case, then pointed a floor-fan (yup, one of those two-foot
> diameter ones) right at the motherboard, voila! I'm not saying this is your
> problem, but it would at least heat as the culprit if you still crashed.
>
> Stefan Lucke wrote:
>
> > [EMAIL PROTECTED] wrote:
> > >
> > > Suffering from an unstable system.
> > > K6-400 (stepping 12), Motherboard FIC PA-2013 (VIA MP3),
> > > 256 MB Ram (PC-100), (the board allows to downclock the RAM to 66 Mhz, what I
> > > did),
> > > AGP Matrox G200, 2 SCSI-Controller, EATA-DPT (only Disks)
> > > and ncr53c825 (DDS-3, CDROM ).
> > > RedHat 5.2 Kernel 2.0.36 and I tried as well all 2.2.0-preXX. the last
> > > 2.2.0-pre7ac2.
> > > The system keeps chrashing.
> > >
> > > I am trying to fix the system now since christmas. Getting frustrated....
> > >
> > > Any ideas, suggestions??
> >
> > Even though I�ve got a differnet mother-board (DFI .., AGP Matrox G200,
> > AHA2940 - disks, dds1-dc, cdrom) Kernel paging errors occure sometimes
> > upon loading of the bttv driver.
> >
> > Did you try other FSB clock rates?
> > With my K6-2 400 it is impossible to compile the kernel with FSB 100MHz
> > (no go at 400, 350, 300 - errors: sig 11 - sig 6). Kernel compile is
> > ok at any FSB 95Mhz rate. (428, 380, 333). But now I�ve some trouble
> > compiling the QT library (sig 11 :-(( at 95 MHz FSB). Perhaps its a good
> > idea to replace the processor. Mine has production lable (below voltage
> > infos) A 9849 EPFW.
> >
> > Stefan Lucke
>
> --
> Christian Aasland
> [EMAIL PROTECTED]
> ICQ#: 30268555
> AIM:caasland
------------------------------
From: Kalle Olavi Niemitalo <[EMAIL PROTECTED]>
Subject: Can I extend struct kernel_stat?
Date: 13 Feb 1999 21:14:46 +0200
I'm porting my irqglitch patch
<http://stekt.oulu.fi/~tosi/misc/linux-2.0.33-irqglitch.diff> to
Linux 2.2.1. Among other things, it adds two glitch counters in
/proc/interrupts. I see the usual counters are now kept in
struct kernel_stat kstat which is declared in
<linux/kernel_stat.h>, so I assume that's where I should add mine
too.
Can I extend the structure without worrying about compatibility?
I'm thinking of the following:
unsigned int pswpin, pswpout;
unsigned int irqs[NR_CPUS][NR_IRQS];
+#ifdef CONFIG_8259A_IRQGLITCH
+ unsigned int irqglitches[NR_CPUS][2];
+#endif
unsigned int ipackets, opackets;
unsigned int ierrors, oerrors;
Would it be better to add the new member at the end of the
structure?
Also, should I add #include <linux/config.h> in the header, or
can I assume users will include it?
------------------------------
From: [EMAIL PROTECTED] (Garson Cole)
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: Sat, 13 Feb 1999 18:25:59 GMT
In article <z%st2.374$[EMAIL PROTECTED]>,
Paul E. Lehmann <[EMAIL PROTECTED]> wrote:
>>Sorry folks - Unix and Linux are about 20 years behind times and will NEVER
>>catch up.
Gee...is that why NT supports IPv6? Oh wait...it doesn't. It's built
into the linux kernel, though.
------------------------------
From: [EMAIL PROTECTED] (Craig J Copi)
Subject: Re: Problem with autofs and local /home
Date: 13 Feb 1999 20:20:06 GMT
In article <7a2oiu$lj2$[EMAIL PROTECTED]>,
[EMAIL PROTECTED] (H. Peter Anvin) writes:
> Followup to: <7a2iea$igs$[EMAIL PROTECTED]>
> By author: [EMAIL PROTECTED] (Craig J Copi)
> In newsgroup: comp.os.linux.development.system
>>
>> I am using the latest autofs with kernel 2.2.1. The users home directory
>> actually lives in /export/home by I use autofs to "mount" this in /home
>> (/export/home is then nfs exported and autofs on clients map them to /home
>> also). On external clients it works great. On local clients it makes a
>> symbolic as it should so
>> /home/test -> /export/home/test
>> Now I do the following
>> root> cd /home/test
>> root> pwd
>> /home/test
>>
>> This is good.
>>
>> root> su - test
>> test> pwd
>> /export/home/test
>>
>> This is bad! This breaks scripts (or GNU queue) that do things like
>> rsh remotehost "cd `pwd`; ./runbinary_in_this_dir"
>> I don't know why the two cases above are giving different results.
>>
>> Any help would be appreciated.
>>
>
> "pwd" is a bash built-in, so it knows how you got there. "su", or "rsh"
> don't have that information.
>
> -hpa
So does this mean I'm stuck? On the nfs server I have to keep the
home directories in /home, export them from here, and not use autofs if I
want to maintain similar looking directory structures across multiple
machines? Is there a simple work around for this that I'm missing? This
would seem like a big problem if I were using nis maps for autofs.
Craig
--
=================================+====================================
Craig J Copi | [EMAIL PROTECTED]
Case Western Reserve University | http://erebus.phys.cwru.edu/~copi/
Department of Physics | (216) 368-8831
------------------------------
From: Ricky Nelson <[EMAIL PROTECTED]>
Subject: graphics lib?
Date: Sat, 13 Feb 1999 17:05:40 -0500
This is a multi-part message in MIME format.
==============D97A6CC3253258DE2C037C45
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
I am new to Linux programming and I am trying to find a standard way of
controlling the display in text mode using C++. Currently I use MS
v1.52 compiler for DOS and it has a graphics library containing
functions like:
_settextposition()
_clearscrean()
_setvideomode()
_setbkcolor()
_settextcolor()
etc...
My question is what is the standard graphics library that is used in
Linux. I am using Redhat v5.2.
I looked at svgalib and ncurses but I couldn't seem to find a way to
move the cursor around the way I am used to. Also is union REGS just a
DOS way of accessing the video display? I couldn't seem to find any
documentation on this in Linux either. Please forgive me for my
ignorance, and any help would be greatly appreciated :)
Ricy Nelson
[EMAIL PROTECTED]
==============D97A6CC3253258DE2C037C45
Content-Type: text/x-vcard; charset=us-ascii;
name="ricky.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Ricky Nelson
Content-Disposition: attachment;
filename="ricky.vcf"
begin:vcard
n:Nelson;Ricky
x-mozilla-html:TRUE
adr:;;2180 South Loudoun Street #120;Winchester;VA;22601;USA
version:2.1
email;internet:[EMAIL PROTECTED]
note:ICQ#: 15049905
x-mozilla-cpt:;-1
fn:Ricky Nelson
end:vcard
==============D97A6CC3253258DE2C037C45==
------------------------------
From: [EMAIL PROTECTED] (Michal Szymanski)
Subject: __setfpucw disappeared from glibc
Date: 12 Feb 1999 14:10:21 GMT
In order to install kernel-level NFS daemons in 2.2.1 kernel running
on RedHat 5.2 system, I had to install newest glibc package found
on rawhide.redhat.com development server (glibc-2.0.111-0.990127).
This was because kernel-level NFS daemons were failing to run
on "original" glibc-2.0.7-29 coming with RH5.2.
Unfortunately, although NFS is now running very well, the new glibc
has broken other applications, some of which fail now reporting:
error in loading shared libraries: /iraf/iraf/bin.redhat/cl.e: undefined
symbol: __setfpucw
inspecting the libc.so.6 with 'nm' I found following:
0001d374 t __setfpucw in 2.0.111
0001b310 T __setfpucw in 2.0.7
which means that __setfpucw, though defined, is now "local" symbol instead of
"global".
I'd appreciate any help on this problem. It is unfortunate, as I need
these applications badly and, at the same time, I would not like to
get rid of kernel-level NFS which so much better!
regards, Michal.
--
Michal Szymanski ([EMAIL PROTECTED])
Warsaw University Observatory, Warszawa, POLAND
------------------------------
From: [EMAIL PROTECTED] (Juergen Heinzl)
Subject: Re: @@GLIBC
Date: Sat, 13 Feb 1999 22:23:00 GMT
In article <[EMAIL PROTECTED]>, Dan Shechter wrote:
>Hi,
>I have a Redhat 5.1 system + Glibc system
>Recently I've noticed that all kinds of RPM's aren't functioning
>properly because they fail to find symbols from libc.so, which
>have a "wierd" @@GLICC2.0 suffix...
Symbol versioning ...
>Can someone explain what is the meaning of this?
... it will help to keep things compatible even if the library interfaces
change. Means assume xbill is compiled against glibc-2.1, now glibc-2.2
introduces an interface change due to added kernel functionality, standards
requirement or whatever. The solution here is to keep the old interface
and to translate to the new one.
[...]
>The current version of GLIBC I have, has no @@GLIBC2.0, should I get
>a newer one?, If so, which? Are the older symbols still supported, e.g.,
glibc-2.1 is the current one, not available though right now.
Cheers,
Juergen
--
\ Real name : J�rgen Heinzl \ no flames /
\ EMail Private : [EMAIL PROTECTED] \ send money instead /
\ Phone Private : +44 181-332 0750 \ /
------------------------------
** 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
******************************
