Linux-Development-Sys Digest #654, Volume #6 Tue, 27 Apr 99 02:14:11 EDT
Contents:
regexp-able syslogd ("Stefan Monnier "
<[EMAIL PROTECTED]>)
calculate absolute sector number (Peter Daum)
syslogging of mountd and autofs ("Stefan Monnier "
<[EMAIL PROTECTED]>)
exhausted memory ([EMAIL PROTECTED])
Re: Compiling 2.2.6 kernel on Slackware 96 (George Domenikiotis)
Re: Get caller's address of a telnet connection (ellis)
Re: exhausted memory (ellis)
Re: Linux disk defragmenter (Mark Hahn)
Help on Compiling kernel 2.2.6 (Simon Su)
Re: redhat 6.0? ("Bob Taylor")
problems configuring an apache server ("JB")
Re: Change from gcc to egcs gives seg fault in make (Clifford T. Matthews)
Re: Help on Compiling kernel 2.2.6 (Nathan Myers)
Malloc-ing out of specific memory (Marcin Romaszewicz)
Re: redhat 6.0? (bryan)
Problem installing makemake with libstdc++ (Fred Dunn)
Re: Malloc-ing out of specific memory (Scott Lanning)
----------------------------------------------------------------------------
From: "Stefan Monnier <[EMAIL PROTECTED]>"
<[EMAIL PROTECTED]>
Crossposted-To: comp.security.unix
Subject: regexp-able syslogd
Date: 26 Apr 1999 15:01:58 -0400
Is there a syslogd replacement that can do pattern matching to
decide where to send a given message (rather than being limited
to severity/facility) ?
I know I can use pipes as message destinations and finetune the
filtering outside syslogd, but it would seem so generally useful
and so much less bug-prone to have such a feature built into syslogd.
Stefan
------------------------------
From: [EMAIL PROTECTED] (Peter Daum)
Crossposted-To: comp.os.linux.development.apps
Subject: calculate absolute sector number
Date: 26 Apr 1999 20:26:39 GMT
Reply-To: [EMAIL PROTECTED]
Hi,
Is there any easy way for a program (C or Perl) to find out the
absolute position (counted in sectors) at which a certain disk
partition starts? example:
> Disk /dev/sdb: 70 heads, 62 sectors, 1015 cylinders
> Units = cylinders of 4340 * 512 bytes
>
> Device Boot Start End Blocks Id System
> /dev/sdb1 1 444 963449 83 Linux native
> /dev/sdb2 445 600 338520 6 DOS 16-bit >=32M
The size of one cylinder here is 4340 sectors, so sdb2 starts on sector:
444*4340=1926960
If nothing else helps, I could of course parse the output from fdisk and
calculate it from that information, but there's got to be a better way!
Any ideas? Help would be appreciated!
regards,
Peter
--
__o
Peter Daum <gator at cs.tu-berlin.de> _'\<_
- pgp messages welcome - ____(_)/(_)
------------------------------
From: "Stefan Monnier <[EMAIL PROTECTED]>"
<[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.admin
Subject: syslogging of mountd and autofs
Date: 26 Apr 1999 16:56:08 -0400
Am I the only one that's annoyed by the syslog messages of mountd
and autofs ?
Why do they have to put and mount/unmount with severity NOTICE ?
These are not unusual events at all, so they should at most be logged
with severity INFO.
I have the impression that mountd was doing the right thing until
I switched to RedHat-5.2. Has something changed there ?
Is there a way to convince mountd and autofs to be less verbose?
Stefan
------------------------------
From: [EMAIL PROTECTED]
Subject: exhausted memory
Date: Mon, 26 Apr 1999 21:17:20 GMT
Is there a way to free up already used memory?
I really hate to reboot after compiling many programs because there is no more
free memory, I'm always left with about 1k free. Can anyone help me find a
solution. Thanks a lot.
============= Posted via Deja News, The Discussion Network ============
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
------------------------------
From: George Domenikiotis <[EMAIL PROTECTED]>
Subject: Re: Compiling 2.2.6 kernel on Slackware 96
Date: Tue, 27 Apr 1999 01:23:27 +0300
Thank you very much Robert and everyone who answered my requests. My problem
was due to versions inconsistencies i.e. older programs than the ones
described in Documentation/Changes. With help like that no wonder people like
me are trying to totally migrate to Linux. The best advocacy Linux can "buy".
My best regards - George
Robert B. Hamilton wrote:
> > Hello everybody,
>
> Hello.
>
> > I have the foll problem: I've compiled the new 2.2.6 and get various
> > unresolved symbols and also the foll message during boot:
> > "kernel version needed but not found" or something like that.
>
> Hm, this looks like a modules error message if my memory serves correctly.
> Perhaps you left out "make modules" and make "modules install", with
> the loadable module support and kernel module loader selected?
>
> > Does the 2.2.x kernel need a newer libc version?
>
> The kernel itself doesnt use libc at all. Some utilities used by the
> kernel build will link to libc, however. The Documentation/Changes
> gives a reccomended version for libc. If this were a problem though,
> you would see some nasty error messages on the configure/build....
>
> > Is there a need to compile as bzImage instead of zImage?
>
> Not necessarily. If there were a need, then there would be an
> error message telling you so.
> --Robert
--
My computer needed upgrade. So, I installed Linux.
------------------------------
From: [EMAIL PROTECTED] (ellis)
Subject: Re: Get caller's address of a telnet connection
Date: 26 Apr 1999 21:58:40 GMT
In article <[EMAIL PROTECTED]>,
Francois Desarmenien <[EMAIL PROTECTED]> wrote:
>I wonder if I there is a way to get the IP address of
>the client which made a connection through telnet.
>
>The problem is I'm in a running process and I know
>it's through a telnet connection. At some point in the
>process, I need to connect to a service on the system
>that connected through telnet.
>
>Can I get this information with the pseudo-tty I'm on ?
>
>Does someone have any clues on how I could do that ?
Have you looked at utmp?
--
http://www.fnet.net/~ellis/photo/linux.html
------------------------------
From: [EMAIL PROTECTED] (ellis)
Subject: Re: exhausted memory
Date: 26 Apr 1999 21:55:58 GMT
In article <7g2l4r$5je$[EMAIL PROTECTED]>, <[EMAIL PROTECTED]> wrote:
>Is there a way to free up already used memory?
What makes you think it isn't freed up?
>I really hate to reboot after compiling many programs because there is no more
>free memory, I'm always left with about 1k free.
Probably more like 1 Meg. But isn't it used in cache and buffers? If so,
why worry about it?
>Can anyone help me find a solution. Thanks a lot.
I doubt you actually have a problem.
--
http://www.fnet.net/~ellis/photo/linux.html
------------------------------
From: Mark Hahn <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.advocacy
Subject: Re: Linux disk defragmenter
Date: 26 Apr 1999 18:39:33 GMT
>>My understanding of disk fragmentation is that the clusters containing
>>the data in, say, a file are not contiguous on the disk surface; such as
>>when a file is modified and enlarged many separeate times, each time
>>beyond its original "footprint" on the disk. Surely this happens with
>>ext2fs as with FAT16/32?
basically no. sure, it can happen if you have a fairly full FS,
and do the writes of certain sizes, interleaved with other allocations.
there are several heuristics that make ext2 resistant to this, though:
one is that it tries to preallocate ~8 blocks ahead, and that it is
smart about using contig spans in the free-block bitmap.
> b) Nobody knows anymore whether data is actually being plased
> continuously on the disk anyways.
that's not really true. ignoring bad-block sparing, linear disk
addresses are contiguous, in the sense of sectors, heads then cyls
filling fastest. that might not be optimal, but it's close.
> disk and thereby optimizing the scheduling of accesses; Linus himself
> has been on record as saying that this is pointless because smart new
> disks prevent you from knowing physical locations.
Linus has a tendency to state things with force ;)
"pointless" is clearly too strong. but it's also true that you
don't/can't know everything about layout. you don't need to though:
a filesystem that tries to keep "access locality" blocks in linearly
contiguous addresses will do well.
> c) Some degree of fragmentation is inherent in having multiple file
> streams active concurrently.
Ext2, like BSD FFS before it, does a certain amount of spatial
demultiplexing of multiple file streams, normally based on parent dir.
all that said, I normally see Ext2 delivering very near platter-level
performance on virgin or at least clean disks; degrading to around .7
performance on FS's that have been multiply abused over years...
regards, mark hahn.
--
operator may differ from spokesperson. [EMAIL PROTECTED]
http://java.mcmaster.ca/~hahn
------------------------------
From: Simon Su <[EMAIL PROTECTED]>
Subject: Help on Compiling kernel 2.2.6
Date: Mon, 26 Apr 1999 17:17:35 -0500
Hello,
I have the following error when I compile the kernel on a PIII system.
make -C arch/i386/lib
make[1]: Entering directory `/usr/src/linux-2.2.6/arch/i386/lib'
make all_targets
make[2]: Entering directory `/usr/src/linux-2.2.6/arch/i386/lib'
gcc -D__KERNEL__ -I/usr/src/linux-2.2.6/include -Wall
-Wstrict-prototypes
-O2 -fomit-frame-pointer -D__SMP__ -pipe -fno-strength-reduce -m486
-malign-loops=2 -malign-jumps=2 -malign-functions=2 -DCPU=686 -c -o
checksum.o checksum.c
checksum.c:200: redefinition of `csum_partial_copy'
checksum.c:105: `csum_partial_copy' previously defined here
{standard input}: Assembler messages:
{standard input}:185: Fatal error: Symbol csum_partial_copy already
defined.
make[2]: *** [checksum.o] Error 1
make[2]: Leaving directory `/usr/src/linux-2.2.6/arch/i386/lib'
make[1]: *** [first_rule] Error 2
make[1]: Leaving directory `/usr/src/linux-2.2.6/arch/i386/lib'
make: *** [_dir_arch/i386/lib] Error 2
Any idea how do get around the probelm? Greatly appreciate any tips that
you have. Thanks a lot
Simon
------------------------------
Reply-To: [EMAIL PROTECTED]
From: [EMAIL PROTECTED] ("Bob Taylor")
Subject: Re: redhat 6.0?
Crossposted-To: comp.os.linux.advocacy
Date: Mon, 26 Apr 1999 23:05:20 GMT
In article <[EMAIL PROTECTED]>,
"Bobby D. Bryant" <[EMAIL PROTECTED]> writes:
> [EMAIL PROTECTED] wrote:
>
>> Any word from RedHat when on the official release date for RedHat linux 6.0
>> is going to be?
>
> Earlybird Alert: Slashdot is reporting that it is already hitting the mirrors.
Sad. Although GNOME is usable, as of this past weekend, there are some show
stopper bugs.
Bob
--
+----------------------------------------------------------------+
| Bob Taylor Email: [EMAIL PROTECTED] |
|----------------------------------------------------------------|
| Gnome certainly is (serious competition to the Mac or Windows) |
| ... I get a charge out of seeing the X Window System work the |
| way we intended..." - Jim Gettys |
+----------------------------------------------------------------+
------------------------------
From: "JB" <[EMAIL PROTECTED]>
Subject: problems configuring an apache server
Date: Mon, 26 Apr 1999 18:45:33 +0100
Hi,
I actually try to configure an apache server on linux to support ASP
capabilities. To do so I tried first to load the mod_perl module with my
apache config but simply loading it doesn't seem to work and I can't reach
compiling it as a share module. molding it directly in apcache compilation
as an supplementary module doesn't seem to work either.
Could someone please tell me what steps to go through or send me references
of a good documentation ?
Thanks for your help,
Max.
------------------------------
From: [EMAIL PROTECTED] (Clifford T. Matthews)
Subject: Re: Change from gcc to egcs gives seg fault in make
Date: 26 Apr 1999 17:34:13 -0600
>>>>> "Stefan" == Stefan Skopnik <[EMAIL PROTECTED]> writes:
Stefan> Hi, I updated from gcc 2.7.2 to egcs 1.1.? on my SuSe 6.1
Stefan> system via rpm packages.
Stefan> Now I get sporadic Segmentation faults from make when
Stefan> compiling the kernel and other software packages.
Stefan> I'm sure it worked before: I compiled a number of kernels
Stefan> successfully with gcc.
Stefan> Now I have to restart 'make bzImage' a dozend times before
Stefan> the kernel compile is finished.
Stefan> The errors definitly come from the make tool not the compiler
Stefan> and mostly when make changed to another directory and
Stefan> restarts again
I doubt that what I'm about to describe is the source of your trouble,
but it was definitely the source of some trouble over here and it was
a bit tricky to track down. It might apply to you.
We do most of our development on Red Hat distributions. We compile
our libc5 binaries on a Red Hat 4.2 system and our glibc binaries on a
Red Hat 5.2 system. However, our glibc binaries (the application
itself, not the compiling of it) were dying on SuSE 6.0 and the beta
of Red Hat 6.0. Eventually I tracked it down to a change in how the
compiler returns structs that are larger than 32-bits. This change
apparently happened when Red Hat switched from gcc to egcs, although
it may be a simple matter of using different configuration options to
the compiler.
This bit us because of our use of the ndbm routines present in the db
shared library. They return a struct that has two pointers in it.
Both compilers make room for this struct on the stack, then pass an
additional pointer to the struct-returning function. The difference
is that under gcc (where we were compiling our code), the caller is
expected to pop that additional pointer. Under egcs (which is called
gcc), the callee is expected to pop that additional pointer.
So if the caller and callee are compiled using the two different
conventions, the stack can get out of sync. I say "can" rather than
"will" because if the compiler is deferring pops, there's a chance
that everything will work out OK, since accesses to local variables
via the frame are still correct and the stack pointer will be returned
to normal when the "leave" instruction is hit (the fact that you're
returning a struct > 32 bits forces a frame pointer).
But if you're not deferring pops, the callee may adjust the stack and
then have the caller adjust the stack, so the stack pointer is now
4-bytes beyond where it should be. If you take an interrupt at this
point (before the temporary data has been copied to where it's
needed), the temporary space that is being used to return the struct
will get clobbered, but if you don't take the interrupt and don't
really count on the stack pointer being correct, you may make it to
the leave instruction without any damage.
Now, I don't think that's what's biting you, because in all likelihood
the make executable was compiled with the same calling conventions as
its shared-libraries were. Even if make had been compiled under gcc
and it uses shared-libraries that were compiled under egcs, it would
only be a problem if these shared libraries returned structs greater
than 32 bits, and that's fairly uncommon (although ndbm definitely
does this).
Still, since the subject fits the pattern ".*Change from gcc to egcs
gives seg fault in.*", perhaps someone who is being bitten by this bug
will read this message and not spend quite as much time tracking it
down as I did.
Regards,
Cliff Matthews <[EMAIL PROTECTED]>
------------------------------
From: [EMAIL PROTECTED] (Nathan Myers)
Subject: Re: Help on Compiling kernel 2.2.6
Date: 26 Apr 1999 17:38:49 -0700
Simon Su <[EMAIL PROTECTED]> wrote:
> I have the following error when I compile the kernel on a PIII system.
>checksum.o checksum.c
>checksum.c:200: redefinition of `csum_partial_copy'
>checksum.c:105: `csum_partial_copy' previously defined here
This happened because you unpacked the kernel tarball right
on top of the old source tree. Bad idea. Some people suggest
deleting checksum.c, but it would be wiser to delete the whole
tree and start over.
--
Nathan Myers
[EMAIL PROTECTED] http://www.cantrip.org/
------------------------------
From: marcin@asmodean (Marcin Romaszewicz)
Subject: Malloc-ing out of specific memory
Date: 27 Apr 1999 00:45:26 GMT
Reply-To: [EMAIL PROTECTED]
Hello all,
I need a malloc which will function on a specific memory area, a big plus
if it's multiprocess safe. I need to set up a shared memory area between a
set of processes, and malloc out of that memory area.
Before I go re-inventing the wheel, I wonder if there already is some
kind of modified malloc that would do this.
Thanks,
-- Marcin
------------------------------
From: bryan <[EMAIL PROTECTED]>
Subject: Re: redhat 6.0?
Crossposted-To: comp.os.linux.advocacy
Date: Tue, 27 Apr 1999 00:52:37 GMT
In comp.os.linux.development.system "Bob Taylor" <[EMAIL PROTECTED]>
wrote:
: In article <[EMAIL PROTECTED]>,
: "Bobby D. Bryant" <[EMAIL PROTECTED]> writes:
: > [EMAIL PROTECTED] wrote:
: >
: >> Any word from RedHat when on the official release date for RedHat linux 6.0
: >> is going to be?
: >
: > Earlybird Alert: Slashdot is reporting that it is already hitting the mirrors.
: Sad. Although GNOME is usable, as of this past weekend, there are some show
: stopper bugs.
that's why anyone who wants stability goes for x.y where y != 0
--
Bryan
------------------------------
From: [EMAIL PROTECTED] (Fred Dunn)
Crossposted-To: linux.redhat.devel,comp.os.linux.development.apps
Subject: Problem installing makemake with libstdc++
Date: Tue, 27 Apr 1999 01:09:54 GMT
Gak! I'm new to this whole Linux development game and I'm stuck. I'm
attempting to install makemake-2.5.0-1.i386.rpm on a stock Redhat 5.2
system and here's what I get as output:
[root@Banff /tmp]# rpm -ivh makemake*
error: failed dependencies:
libstdc++-libc6.1.1.so.2 is needed by makemake-2.5.0-1
[root@Banff /tmp]# rpm -q libstdc++
libstdc++-2.8.0-14
[root@Banff /tmp]# find / -name libstdc* -mount
/usr/lib/libstdc++.so.2.7.2.8
/usr/lib/libstdc++.so.2.7.2
/usr/lib/libstdc++.so.2.8.0
/usr/lib/libstdc++.so.2.8
/usr/lib/libstdc++.a
/usr/lib/libstdc++.so
/usr/i486-linux-lib5/lib/libstdc++.so.27.1.4
/usr/i486-linux-lib5/lib/libstdc++.so.27
/usr/glibc/libstdc++.so.2.72.8
It seems to be there and I've set LD_LIBRARY_PATH to
/usr/lib:/usr/local/bin:/bin:/usr/bin:/usr/X11R6/bin
So, anyone know what gives? I'm stuck until I solve this so any help
is appreciated.
------------------------------
From: [EMAIL PROTECTED] (Scott Lanning)
Subject: Re: Malloc-ing out of specific memory
Date: 27 Apr 1999 01:50:53 GMT
Marcin Romaszewicz (marcin@asmodean) wrote:
: I need a malloc which will function on a specific memory area, a big plus
: if it's multiprocess safe. I need to set up a shared memory area between a
: set of processes, and malloc out of that memory area.
Maybe you want to look at mmap(). Something like
space = mmap(address, length, PROT_something, MAP_something, fd, offset);
For example, MAP_SHARED would set up a space in a file for processes to
share. Also mlock() locks some memory space.
------------------------------
** 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
******************************
