Linux-Development-Sys Digest #951, Volume #6 Mon, 12 Jul 99 13:14:06 EDT
Contents:
Memory error tool (Barry Fruitman)
Re: After Week 1 With Linux -- licking wounds. (Eric Fowler)
Re: MICROSOFT LINUX DISTRIBUTION (Greg White)
anonymous memory mapping (Takeyasu Wakabayashi)
Re: anonymous memory mapping (Frank v Waveren)
iBCS in 2.2.10 (capai)
unable to post a message bigger than 75ko ("Christophe Pecquerie")
Re: anonymous memory mapping (Mattias Engdeg�rd)
Where is Glibc2.7.2.1 URL (Matt)
Re: how to "mount" a distribution ? (Marcus Sundberg)
Re: iBCS in 2.2.10 (fredrik chabot)
Accessing kernel from user application (Mats Liljegren)
ioctl() / security (Dimitri Papadopoulos)
Re: MICROSOFT LINUX DISTRIBUTION (Christopher B. Browne)
Re: Where is Glibc2.7.2.1 URL (Mark Tranchant)
Re: ioctl() / security (Marius van Wyk (remove NOSPAM.))
Re: ioctl() / security (Dimitri Papadopoulos)
Re: Driver for AMCC 5933 S (Pascal Hirsch)
Re: ioctl() / security (Dimitri Papadopoulos)
Debugging with NMI (David Grothe)
2.2.10/ppp/klogd problem (Peter Baker)
Help: installing glibc 2.1.1 on top of 2.0.7 Runtime (Dr H. T. Leung)
----------------------------------------------------------------------------
From: Barry Fruitman <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.development.apps
Subject: Memory error tool
Date: Mon, 12 Jul 1999 04:38:58 GMT
Hi all,
Can anyone out there recommend a good memory-debugging tool (a la Purify
or Boundschecker) to find memory errors in my gcc programs? I downloaded
GNU's Checker but couldn't even get it to compile on my system with gcc
v2.91.66 (which is higher than the required version). I do believe I am
using glibc2 but how do I check for sure? I'm running Linux-Mandrake
6.0. The version of Checker I found is v0.9.9.1 which is beta anyway.
I there a good (free) memory-debugger out there?
Any and all suggestions are much appreciated.
Kindly reply via email if possible.
Thanks,
Barry Fruitman
------------------------------
From: Eric Fowler <[EMAIL PROTECTED]>
Subject: Re: After Week 1 With Linux -- licking wounds.
Date: Sun, 11 Jul 1999 18:36:01 -0600
Tyler wrote:
> It is absolutely ridiculous that we should spend our
> time going around to each computer with a Netscape disk to install the
> newest version when I could do it on one computer and update the entire
> building in a UNIX installation.
Not to defend Micro$oft, but couldn't you do this with SMS?
--
Vivez sans temps mort!
(Live without dead time)
-Situationist International
sockeye [at] rmii [dot] com
------------------------------
From: Greg White <[EMAIL PROTECTED]>
Subject: Re: MICROSOFT LINUX DISTRIBUTION
Date: Mon, 12 Jul 1999 05:42:29 GMT
SNIP
> - Novell's Netware product provides its own operating system. It does not
> need MS-DOS or Linux or anything else in order to function.
Huh?? Exactly how do you get server.exe to run??? It's a DOS binary...
There may be a way I'm not aware of, but every 3.x and 4.x and 5.x
install I've done starts with "Create a small MS-DOS FAT partition..."
GW
------------------------------
From: Takeyasu Wakabayashi <[EMAIL PROTECTED]>
Subject: anonymous memory mapping
Date: 12 Jul 1999 14:45:30 +0900
This may be a FAQ, but let me ask anyway.
I tried to use anonymous memory mapping to share a memory
region betweeen forked processes. The code goes like this:
addr = (int *)mmap(NULL, sizeof(int), PROT_READ | PROT_WRITE,
MAP_SHARED | MAP_ANON, -1, 0);
if (addr == MAP_FAILED) {
perror("mmap");
exit(1);
}
*addr = 0;
When I run the program, mmap fails with EINVAL.
I also tried /dev/zero anonymous memory mapping and the result
was the same. It seems that Linux doesn't allow the sharing of
anonymously mapped memory regions between processes.
My questions are:
1. I tried on 2.1.129(on Alpha) and 2.0.36(on i386). Does 2.2.x
allow the sharing of anonymously mapped memory regions?
2. Do I have to create a file to map or to use System Five shared
memory to share memory regions? (or clone() or threads...)
3. Is this a deliberate implementation? If so, why?
Thanks in advance,
--
Takeyasu Wakabayashi,
Faculty of Economics, Toyama University
[EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Frank v Waveren)
Subject: Re: anonymous memory mapping
Date: Mon, 12 Jul 1999 09:06:31 GMT
In article <[EMAIL PROTECTED]>,
Takeyasu Wakabayashi <[EMAIL PROTECTED]> writes:
> This may be a FAQ, but let me ask anyway.
I don't know the answer to this, but it's not a FAQ. I WISH IT WAS! :-)
--
Frank v Waveren
[EMAIL PROTECTED]
ICQ# 10074100
------------------------------
From: capai <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc,comp.os.linux.setup
Subject: iBCS in 2.2.10
Date: 12 Jul 1999 08:30:58 GMT
in 2.2.10 kernel.
i am trying to install iBCS-2.1-981105 so i ftp(ed) its sources from
ftp.cdrom.com...
compiling iBCS..
make
make install
after installing the compiled iBCS i issued a 'depmod -a' which returned
modprobe: not an ELF
/lib/modules/2.2.10/misc/IBCS: unresolved symbol(s)
why???? how do i solve this???
thnx...
.c.a.p.a.i.
================== Posted via SearchLinux ==================
http://www.searchlinux.com
------------------------------
From: "Christophe Pecquerie" <[EMAIL PROTECTED]>
Subject: unable to post a message bigger than 75ko
Date: Mon, 12 Jul 1999 12:32:39 +0200
We use a linux server for our internal newsgroup server but we can't send
message bigger than 75000 bytes. what's the way to switch the size limit to
2Mbytes or more? Thanks a lot for your help.
------------------------------
From: [EMAIL PROTECTED] (Mattias Engdeg�rd)
Subject: Re: anonymous memory mapping
Date: 12 Jul 99 10:07:10 GMT
Takeyasu Wakabayashi <[EMAIL PROTECTED]> writes:
>I also tried /dev/zero anonymous memory mapping and the result
>was the same. It seems that Linux doesn't allow the sharing of
>anonymously mapped memory regions between processes.
Right; shared writable anonymous mappings are not supported yet.
>1. I tried on 2.1.129(on Alpha) and 2.0.36(on i386). Does 2.2.x
> allow the sharing of anonymously mapped memory regions?
No. Eric Biederman has written patches for this (I've lost the link;
look for "shmfs").
>2. Do I have to create a file to map or to use System Five shared
> memory to share memory regions? (or clone() or threads...)
Yes.
>3. Is this a deliberate implementation? If so, why?
I'll let someone who better knows the MM internals of the kernel answer this.
------------------------------
From: Matt <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.setup
Subject: Where is Glibc2.7.2.1 URL
Date: Mon, 12 Jul 1999 11:48:29 +0100
Hi,
I would like to build glibc2.7.2.1 but where can I download
either the rpm's for SuSE from or the full source code
from ?
Also are there any tips or would the glibc-howto-do be ok ?
Many thanks
Matt
------------------------------
From: Marcus Sundberg <[EMAIL PROTECTED]>
Subject: Re: how to "mount" a distribution ?
Date: Mon, 12 Jul 1999 12:55:35 +0200
Rafael The X-Slasher wrote:
>
> Anyone knpws any howto/document/tutorial on how to "mount"/create a
> linux distribution ?
1. Read the mount(8) man page.
2. Go to http://metalab.unc.edu/LDP/ and read all the documentation you
find. When you have read and understood everything there you hopefully
know enough to make a distribution.
//Marcus
--
===============================+====================================
Marcus Sundberg | http://www.stacken.kth.se/~mackan/
Royal Institute of Technology | Phone: +46 707 295404
Stockholm, Sweden | E-Mail: [EMAIL PROTECTED]
------------------------------
From: fredrik chabot <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc,comp.os.linux.setup
Subject: Re: iBCS in 2.2.10
Date: Mon, 12 Jul 1999 13:29:06 +0200
Reply-To: [EMAIL PROTECTED]
capai wrote:
> in 2.2.10 kernel. i am trying to install iBCS-2.1-981105 so i ftp(ed) its
> sources from ftp.cdrom.com...
Ok I had similar problems, I've compiled the iBCS from the redhat kernel
source witch at least ran.
But i've had al kinds of problems afterwards i.e. kernel panic's,
unexplained stops et al. no problems after dumping iBCS (and replacing the
sco app)
--
Met vriendelijke groet,
Fredrik Chabot
http://www.f6.nl/
------------------------------
From: Mats Liljegren <[EMAIL PROTECTED]>
Subject: Accessing kernel from user application
Date: Mon, 12 Jul 1999 13:52:02 +0200
Hi all,
I've done some experimenting with /usr/include/asm/unistd.h, and for a
statically linked application that is good stuff! As it is now, I have a
statically linked application being 17KB big using that interface, while
it was around 120KB big using libc instead...
However, there are (of course) some functionality I miss, even though I
do simple applications. For example, dynamic memory handling. Are there
other interfaces a user application can use to make the most out of the
kernel, apart from the unistd.h?
/Mats
------------------------------
From: Dimitri Papadopoulos <[EMAIL PROTECTED]>
Subject: ioctl() / security
Date: Mon, 12 Jul 1999 13:43:11 +0200
Hi,
Trying to eject CD-ROMs and Jaz disks with eject-2.0.2, I came across
strange discrepancies in the way ioctl() calls are handled:
1) I need read permission on the device to eject CD-ROMs (IDE or SCSI):
$ ls -l /dev/scd0
brw-rw-r-- 1 root disk 11, 0 May 5 1998 /dev/scd0
$ eject -r /dev/scd0
$
This won't work without read permissions:
$ ls -l /dev/scd0
brw-rw---- 1 root disk 11, 0 May 5 1998 /dev/scd0
$ eject -r /dev/scd0
eject: unable to find or open device for: `/dev/scd0'
$
It sounds OK, although I don't understand the relation between the
read permission and the call to ioctl(). Is documentation on this
issue available somewhere?
2) I must be root (effective or real UID) to eject a Jaz disk at home.
# ls -l /dev/sda4
brw-rw---- 1 root disk 8, 4 May 5 1998 /dev/sda4
# eject -s /dev/sda4
#
Anything else won't work:
$ ls -l /dev/sda4 /usr/bin/eject
brwxrwxrwx 1 root disk 8, 4 May 5 1998 /dev/sda4
-rwxr-s--x 1 root disk 17404 Feb 17 16:36 /usr/bin/eject
$ eject -s /dev/sda4
eject: unable to find or open device for: `/dev/sda4'
$
I have been told that this is the correct behaviour. Why such a
difference between Jaz and CD-ROMs? I had a look at the eject-2.0.2
sources and I understand ioctl() is called with different arguments
in either case. Still, the intent is the same, so why does ioctl()
react differently?
I am using RedHat 5.1 with a 2.0.34 kernel.
The hardware is an Adaptec AHA-2940A Ultra SCSI host adapter.
3) At office, anyone can eject Jaz disks! We don't even need read
permissions on the device driver:
$ ls -l /dev/sdb4 /usr/bin/eject
brw-rw---- 1 root disk 8, 4 May 5 1998 /dev/sdb4
-rwxr-xr-x 1 root root 17404 Feb 17 16:36 /usr/bin/eject
$ eject -s /dev/sdb4
$
Is this a bug or just a 'different' normal behaviour? Where are such
security issues handled? Are they handled in a central place in the
kernel, or is it up to the device driver programmer to check UIDs
and file permissions? What's the logic behind the current (apparent?)
chaos?
In this case, we are using redHat 5.1 with a 2.0.35 kernel.
The hardware is a different Adaptec SCSI host adapter which uses a
different driver - don't remember which right now, sorry.
------------------------------
From: [EMAIL PROTECTED] (Christopher B. Browne)
Subject: Re: MICROSOFT LINUX DISTRIBUTION
Reply-To: [EMAIL PROTECTED]
Date: Mon, 12 Jul 1999 12:35:40 GMT
On Mon, 12 Jul 1999 05:42:29 GMT, Greg White <[EMAIL PROTECTED]> posted:
>SNIP
>> - Novell's Netware product provides its own operating system. It does not
>> need MS-DOS or Linux or anything else in order to function.
>
>Huh?? Exactly how do you get server.exe to run??? It's a DOS binary...
>There may be a way I'm not aware of, but every 3.x and 4.x and 5.x
>install I've done starts with "Create a small MS-DOS FAT partition..."
... And this dependancy is similar to the dependancy that Linux requires
a bootloader.
The existence of a bootloader does not mean that once Linux is running,
you still have an active instance of [LILO | MS-DOS + LINLOAD.EXE].
In similar manner, Netware may need a host in order to get loaded,
but once running, is not dependant on that host.
--
"In the long run every program becomes rococo - then rubble."
-- Alan Perlis
[EMAIL PROTECTED] <http://www.hex.net/~cbbrowne/lsf.html>
------------------------------
From: Mark Tranchant <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.setup
Subject: Re: Where is Glibc2.7.2.1 URL
Date: Mon, 12 Jul 1999 12:38:14 +0100
Reply-To: [EMAIL PROTECTED]
2.7.2.1? That's the version number for an old version of gcc. glibc is
at 2.1.1 or thereabouts, I think.
Mark.
Matt wrote:
>
> Hi,
>
> I would like to build glibc2.7.2.1 but where can I download
> either the rpm's for SuSE from or the full source code
> from ?
>
> Also are there any tips or would the glibc-howto-do be ok ?
>
> Many thanks
>
> Matt
------------------------------
From: [EMAIL PROTECTED] (Marius van Wyk (remove NOSPAM.))
Subject: Re: ioctl() / security
Date: Mon, 12 Jul 1999 12:37:08 GMT
On Mon, 12 Jul 1999 13:43:11 +0200, Dimitri Papadopoulos <[EMAIL PROTECTED]> wrote:
> Hi,
>
> Trying to eject CD-ROMs and Jaz disks with eject-2.0.2, I came across
> strange discrepancies in the way ioctl() calls are handled:
>
>
> 1) I need read permission on the device to eject CD-ROMs (IDE or SCSI):
> $ ls -l /dev/scd0
> brw-rw-r-- 1 root disk 11, 0 May 5 1998 /dev/scd0
> $ eject -r /dev/scd0
> $
> This won't work without read permissions:
> $ ls -l /dev/scd0
> brw-rw---- 1 root disk 11, 0 May 5 1998 /dev/scd0
> $ eject -r /dev/scd0
> eject: unable to find or open device for: `/dev/scd0'
> $
> It sounds OK, although I don't understand the relation between the
> read permission and the call to ioctl(). Is documentation on this
> issue available somewhere?
>
It may be that it needs to check the state (can't see why myself.)
>
> 3) At office, anyone can eject Jaz disks! We don't even need read
> permissions on the device driver:
> $ ls -l /dev/sdb4 /usr/bin/eject
> brw-rw---- 1 root disk 8, 4 May 5 1998 /dev/sdb4
> -rwxr-xr-x 1 root root 17404 Feb 17 16:36 /usr/bin/eject
> $ eject -s /dev/sdb4
> $
>
> Is this a bug or just a 'different' normal behaviour? Where are such
> security issues handled? Are they handled in a central place in the
> kernel, or is it up to the device driver programmer to check UIDs
> and file permissions? What's the logic behind the current (apparent?)
> chaos?
Check for suid permissions on eject... kill them. -shrug-
Dom.
------------------------------
From: Dimitri Papadopoulos <[EMAIL PROTECTED]>
Subject: Re: ioctl() / security
Date: Mon, 12 Jul 1999 16:12:16 +0200
I just wrote:
>
> "Marius van Wyk (remove NOSPAM.)" wrote:
> >
> > On Mon, 12 Jul 1999 13:43:11 +0200, Dimitri Papadopoulos <[EMAIL PROTECTED]>
>wrote:
> >
> > [...]
> >
> > > 3) At office, anyone can eject Jaz disks! We don't even need read
> > > permissions on the device driver:
> > > $ ls -l /dev/sdb4 /usr/bin/eject
> > > brw-rw---- 1 root disk 8, 4 May 5 1998 /dev/sdb4
> > > -rwxr-xr-x 1 root root 17404 Feb 17 16:36 /usr/bin/eject
>
> No suid permissions on eject, as you can see.
Sorry, sorry, sorry...
suid _is_ set on eject. I must have looked on the wrong machine???
Let's change the question, why should one be root to eject from SCSI
devices? This is not required for CD-ROMs?
We have to suid eject if we want the users to be able to eject
from the Jaz drive. Is this precisely the point - we don't want
users to be able to eject SCSI backup tapes? But then why doesn't
the same hold for CD-ROMs? CD-ROM writers may be used for backup
purposes. Can a user eject the backup CD-ROM?
------------------------------
From: Pascal Hirsch <[EMAIL PROTECTED]>
Subject: Re: Driver for AMCC 5933 S
Date: Mon, 12 Jul 1999 16:18:05 +0200
Patrick Rust wrote:
>
> Dear Linux-System developer,
>
> I have a little problem !
> How can I write in the Outgoing Mail Box (OMB1..4) and how can I read the
> Incoming Mail Box (IMB1..4) from the Kernel Space?
>
> The source code for my driver is nearly finished, but I can't read or write
> the Operation-Registers!
>
> If you would have some little bit source code for me, I would be very
> pleased.
>
Hi Patrick,
take a look at the driver for the ADLink / NuDAQ ADL7200 data
acquisition board. This board is based on the AMCC 5933. I have
successfully tested this driver with Linux 2.0.36. Maybe you can learn
from the code?
Here is the LSM entry for the driver:
Begin3
Title: Ad-Link 7200 device driver
Version: 0.60
Entered-date: 01JUL98
Description: A device driver for the AdLink 7220 digital I/O device.
Keywords: I/O, acquisizion, driver, laboratory
Author: [EMAIL PROTECTED] (Alessandro Rubini)
Maintained-by: [EMAIL PROTECTED] (Alessandro Rubini)
Primary-site: sunsite.unc.edu
/pub/Linux/science/lab/adl7200-0.60.tar.gz
37kB adl7200-0.60.tar.gz
1kB adl7200-0.60.lsm
Original-site: ftp.systemy.it /pub/develop
Platform: linux, PCI
Copying-policy: GPL
End
Pascal
------------------------------
From: Dimitri Papadopoulos <[EMAIL PROTECTED]>
Subject: Re: ioctl() / security
Date: Mon, 12 Jul 1999 15:47:34 +0200
"Marius van Wyk (remove NOSPAM.)" wrote:
>
> On Mon, 12 Jul 1999 13:43:11 +0200, Dimitri Papadopoulos <[EMAIL PROTECTED]>
>wrote:
>
> [...]
>
> > 3) At office, anyone can eject Jaz disks! We don't even need read
> > permissions on the device driver:
> > $ ls -l /dev/sdb4 /usr/bin/eject
> > brw-rw---- 1 root disk 8, 4 May 5 1998 /dev/sdb4
> > -rwxr-xr-x 1 root root 17404 Feb 17 16:36 /usr/bin/eject
No suid permissions on eject, as you can see.
> > $ eject -s /dev/sdb4
> > $
> >
> > Is this a bug or just a 'different' normal behaviour? Where are such
> > security issues handled? Are they handled in a central place in the
> > kernel, or is it up to the device driver programmer to check UIDs
> > and file permissions? What's the logic behind the current (apparent?)
> > chaos?
>
> Check for suid permissions on eject... kill them. -shrug-
'eject' does not have suid permissions. I checked this yesterday.
I don't have the PC right in front of me, but I'll double check.
This is the whole point of my post. If this is a bug, I can update the
kernel and see what happens. But I just don't know which behaviour is
to be considered 'normal'... There are so many different behaviours
depending on the ioctl arguments and on the device driver, that I don't
know what is correct behaviour anymore. Is all this documented somewhere?
Where are file permissions and UID checked? In a single place somewhere
in the kernel (doesn't seem so at first sight...)?
For your information:
- to eject CD-ROMs, 'eject' uses:
status = ioctl(fd, CDROMEJECT);
- to eject from SCSI devices other than CD-ROMs, 'eject' uses:
/* ... */
scsi_cmd.cmd[0] = ALLOW_MEDIUM_REMOVAL;
/* ... */
status = ioctl(fd, SCSI_IOCTL_SEND_COMMAND, (void *)&scsi_cmd);
/* ... */
scsi_cmd.cmd[0] = START_STOP;
/* ... */
status = ioctl(fd, SCSI_IOCTL_SEND_COMMAND, (void *)&scsi_cmd);
/* ... */
scsi_cmd.cmd[0] = START_STOP;
/* ... */
status = ioctl(fd, SCSI_IOCTL_SEND_COMMAND, (void *)&scsi_cmd);
------------------------------
From: David Grothe <[EMAIL PROTECTED]>
Reply-To: [EMAIL PROTECTED]
Subject: Debugging with NMI
Date: Mon, 12 Jul 1999 11:27:32 -0500
Kernel hackers:
Maybe this is old hat, but it is new to me --
On an ISA bus machine, if you short out the A1 and B1 pins of an ISA
slot you will generate an NMI to the CPU. This interrupts even a
machine that is hung in a loop with interrupts disabled. Used in
conjunction with kgdb <
ftp://ftp.gcom.com/pub/linux/src/kgdb-2.2.6/kgdb-2.2.6.tgz > you can
gain debugger control of a machine that is hung in the kernel! Even
without kgdb the kernel will print a stack trace so you can find out
where it was hung.
The A1/B1 pins are directly opposite one another and the farthest pins
towards the bracket end of the ISA bus socket. You can stick a paper
clip or multi-meter probe between them to short them out.
I had a spare ISA bus to PC104 bus adapter around. The PC104 end of the
board consists of two rows of wire wrap pins. So I wired a push button
between the A1/B1 pins and now have an ISA board that I can stick into
any ISA bus slot for debugger entry.
Microsoft has a circuit diagram of a PCI card at
http://www.microsoft.com/hwdev/DEBUGGING/DMPSW.HTM. If you want to
build one you will have to mail them and ask for the PAL equations.
Nobody makes one comercially.
[THIS TIP COMES WITH NO WARRANTY WHATSOEVER. It works for me, but if
your machine catches fire, it is your problem, not mine.]
-- Dave (the kgdb guy)
------------------------------
From: Peter Baker <[EMAIL PROTECTED]>
Crossposted-To: aus.computers.linux
Subject: 2.2.10/ppp/klogd problem
Date: 12 Jul 1999 15:53:20 GMT
Hi
I recently compiled the 2.2.10 kernel on a pretty stock standard RH6.0
system.
Now when I dial up klogd goes into overdrive taking virtually all of
one CPU (out of two) for itself. This doesnt happen under the standard
2.2.5 kernel that comes with RH6.0
How can I diagnose/rectify the problem
thanks
Peter
=========================================================================
|| Peter Baker email: [EMAIL PROTECTED] ph: 07 3864 1293 (W) ||
|| Mail: School of Mathematics, Queensland University ||
|| of Technology, GPO Box 2434, Brisbane QLD 4001 ||
=========================================================================
------------------------------
From: [EMAIL PROTECTED] (Dr H. T. Leung)
Subject: Help: installing glibc 2.1.1 on top of 2.0.7 Runtime
Date: 12 Jul 1999 15:31:26 GMT
I have an almost new Slackware 4.0 system with the glibc 2.0.7pre runtime support
package installed (it is functional - my latex/rpm/wget are taken from RH 5.2). I
just tried to follow the glibc 2 HOWTO to install 2.1.1 as a test library. So I
did
./configure --enable-add-ons --prefix=/usr/local/glibc21
then "make ; make check"; a couple of ieee maths things flashed across with some
warnings, but make check went to the end. then I "make install" and symbolic
link ld-linux.so.2 to /lib ; but after ldconfig , it reverted back to linking to
/lib/ld-2.0.7.so ; I tried hiding ld-2.0.7.so and running ldconfig again. Then I
do as told to create /usr/local/lib/gcc-lib/i686-pc-linux-gnulibc2/ along side
the original spec ( /usr/local/lib/gcc-lib/i686-pc-linux-gnulibc1 - note the 1 at
the end), and edited the spec accordingly, and updated the header files links.
Then I do:
gcc -bi686-pc-linux-gnulibc2 -nostdinc
-I/usr/local/glibc21/include
-I/usr/local/lib/gcc-lib/i686-pc-linux-gnulibc2/egcs-2.91.66/include
glibc.c
Where glibc.c is simply the hello world program.
The resulting binary shows:
ldd a.out
libc.so.5 => not found
Why?
Here is the same command line above with the -v witch added.
Reading specs from /usr/local/lib/gcc-lib/i686-pc-linux-gnulibc2/egcs-2.91.66/specs
gcc version egcs-2.91.66 19990314 (egcs-1.1.2 release)
/usr/local/lib/gcc-lib/i686-pc-linux-gnulibc2/egcs-2.91.66/cpp -lang-c -nostdinc -v
-I/usr/local/glibc21/include
-I/usr/local/lib/gcc-lib/i686-pc-linux-gnulibc2/egcs-2.91.66/include -undef
-D__GNUC__=2 -D__GNUC_MINOR__=91 -D__ELF__ -Dunix -Di386 -Dlinux -D__ELF__ -D__unix__
-D__i386__ -D__linux__ -D__unix -D__i386 -D__linux -Asystem(unix) -Asystem(posix)
-Acpu(i386) -Amachine(i386) -D__i486__ glibc.c /tmp/ccOSwteb.i
GNU CPP version egcs-2.91.66 19990314 (egcs-1.1.2 release) (i386 Linux/ELF)
#include "..." search starts here:
#include <...> search starts here:
/usr/local/glibc21/include
/usr/local/lib/gcc-lib/i686-pc-linux-gnulibc2/egcs-2.91.66/include
End of search list.
/usr/local/lib/gcc-lib/i686-pc-linux-gnulibc2/egcs-2.91.66/cc1 /tmp/ccOSwteb.i -quiet
-dumpbase glibc.c -version -o /tmp/ccaq2hge.s
GNU C version egcs-2.91.66 19990314 (egcs-1.1.2 release) (i686-pc-linux-gnulibc1)
compiled by GNU C version egcs-2.91.57 19980901 (egcs-1.1 release).
as -V -Qy -o /tmp/ccE3IDcj.o /tmp/ccaq2hge.s
GNU assembler version 2.9.1 (i386-pc-linux-gnulibc1), using BFD version 2.9.1.0.19
/usr/local/lib/gcc-lib/i686-pc-linux-gnulibc2/egcs-2.91.66/collect2 -m elf_i386
-dynamic-linker /usr/local/glibc21/lib/ld-linux.so.2 /usr/lib/crt1.o /usr/lib/crti.o
/usr/local/lib/gcc-lib/i686-pc-linux-gnulibc2/egcs-2.91.66/crtbegin.o
-L/usr/local/lib/gcc-lib/i686-pc-linux-gnulibc2/egcs-2.91.66 -L/usr/local/lib
/tmp/ccE3IDcj.o -lgcc -lc -lgcc
/usr/local/lib/gcc-lib/i686-pc-linux-gnulibc2/egcs-2.91.66/crtend.o /usr/lib/crtn.o
--
--------------------------------------------------
"What you don't care cannot hurt you." Chap. 7a, AMS-NS
------------------------------
** 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
******************************
