Linux-Development-Sys Digest #33, Volume #7 Mon, 9 Aug 99 07:14:00 EDT
Contents:
Re: My first linux program: non-bios boot loader (Neil Koozer)
ppp autodial (Robin Becker)
How to call internal/external script in RPM install time ("robert_c")
Re: no-lilo problem (Nathan Myers)
Re: How to call internal/external script in RPM install time (M. David Allen)
Re: no-lilo problem (Nathan Myers)
I can't make zImage with "make zImage" ("BongShin Choi")
I can't make zImage. ("BongShin Choi")
as86/ld86? (Chris Gregory)
Re: Device driver programming and C++ (Eugene Morozov)
"Domain validation" for reading tape archive??? (Josh Stern)
DIGITAL output! Matrox G200. (Vladimir)
printk ??? (Villy Kruse)
Re: The beginnings of a protected-mode boot loader (Etienne Lorrain)
Re: Device driver programming and C++ ("Andrey Fisunenko")
Re: Toshiba DVD-RAM and Linux (Christoph Martin)
Re: My first linux program: non-bios boot loader (Kees J Bot)
Re: My first linux program: non-bios boot loader (Etienne Lorrain)
Re: My first linux program: non-bios boot loader (H. Peter Anvin)
Re: bzip2 compressed kernel (Martin Boening)
----------------------------------------------------------------------------
From: Neil Koozer <[EMAIL PROTECTED]>
Subject: Re: My first linux program: non-bios boot loader
Date: Sun, 08 Aug 1999 15:47:53 +0000
> [...]
> >I think so, but then I don't know why NT can't handle lilo being in the
> >MBR. I would assume that nuni would work in /hda1 (for example) the same as
> >lilo for NT users.
>
> The LILO code doesn't by any chance start with a jump instruction, does
> it? I had problems with NT no longer booting if the Minix (yes, the
> little teaching O.S.) master bootstrap was installed into the MBR.
>
> If the first instruction of the MBR is a short jump then NT will hang at
> boot telling it can't find it's boot file system. My guess is that if
[...]
Hey, this is interesting... LILO (version 20) begins with:
cli
jmp start
and LILO (version 19) begins with:
jmp start
A comment on the 'cli' line say "NT blows up without this"
This would explain how someone recently reported that he put LILO in the MBR
without affecting NT.
Neil.
------------------------------
From: Robin Becker <[EMAIL PROTECTED]>
Subject: ppp autodial
Date: Sun, 8 Aug 1999 06:18:23 +0100
I'm trying to configure RH6.0 to handle autodialup via ppp. The ppp
devices appear to be configured ok and I can use ifup ppp0 to get the
connection up. I would like to get a ping to an outside system to bring
up ppp0 if required. In my old system this was handled by request-route.
Is it not possible with the newer kernel and net code? I have a lot of
scripts to alter if this cannot be done.
--
Robin Becker
------------------------------
From: "robert_c" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.development.apps,comp.os.linux.hardware
Subject: How to call internal/external script in RPM install time
Date: 9 Aug 1999 01:54:47 GMT
Hi:
I have a problem about writing RPM. Couls you please give me any your
advices? Thanks!!!
Now, I want to use shell script (it need to be called in RPM install-time,
not RPM built-time) to get running machine kernel version when I issue
"rpm -ivh XXXX.i386.rpm".
For above purpose, I write my XXXX.spec( one of rpm file) like following.
Which section will be useful (if I have a script, called get_info, it want
to collect running kernel version to decide which directory I will copy
into from XXXX.i386.rpm)? Is it %pre or %built or %install or %post sections
and how to write it in XXXX.spec?
I have tried to insert my script in (%pre, %built and %install) three
sections, but the script just take effective in built machine (in other
words, it take effective in compile time, not RPM install time). The *only*
useful section is %post%. But, if I insert my script in *% post* section,
it is too late. Because I need this infomation (to decide which directory of
running machine I will cp into.) before RPM execute install section.
Thanks for any your help!!
==================================
XXXX.spec
%pre
(Question: insert my script here and how to write it?)
%built
%install
install 755 MyDriver.o /lib/modules/**$KERNEL_VERSION**
%post
------------------------------
From: [EMAIL PROTECTED] (Nathan Myers)
Subject: Re: no-lilo problem
Date: 8 Aug 1999 19:10:48 -0700
Horst von Brand <[EMAIL PROTECTED]> wrote:
> J. Escalante <[EMAIL PROTECTED]> wrote:
>
>>Since I don't have any other OS than Linux Slackware running in my
>>computer, what I want is not to have lilo at boot up time.
>
>Not possible. The linux kernel can't boot by itself (too big), it needs some
>helper to load it. This may be lilo, syslinux, grub, even ntldr or the
>Partition Magic bootloader (the last two AFAIU). If you don't want lilo's
>prompt, check the lilo configuration. OTOH, lilo allows you to select among
>several linux configurations...
Horst is not strictly correct. John Reiser posted a boot block
which understands the ext2 file system and can load linux all by
itself, without a second-stage boot loader. Try a search on
DejaNews or Freshmeat.
--
Nathan Myers
[EMAIL PROTECTED] http://www.cantrip.org/
------------------------------
From: [EMAIL PROTECTED] (M. David Allen)
Crossposted-To: comp.os.linux.development.apps,comp.os.linux.hardware
Subject: Re: How to call internal/external script in RPM install time
Date: 9 Aug 1999 04:01:24 GMT
Reply-To: [EMAIL PROTECTED]
In article <7olcd7$cf6$[EMAIL PROTECTED]>,
"robert_c" <[EMAIL PROTECTED]> writes:
> Hi:
>
> I have a problem about writing RPM. Couls you please give me any your
> advices? Thanks!!!
>
> Now, I want to use shell script (it need to be called in RPM install-time,
> not RPM built-time) to get running machine kernel version when I issue
> "rpm -ivh XXXX.i386.rpm".
There's a flag to uname that will give you kernel version...I just don't
remember what it is. Check the manpage.
--
David Allen
http://opop.nols.com/
========================================
Yogi Berra, upon hearing his wife say she had been to see
"Doctor Zhivago," said, "What's the matter with you _now_?"
------------------------------
From: [EMAIL PROTECTED] (Nathan Myers)
Subject: Re: no-lilo problem
Date: 8 Aug 1999 19:15:18 -0700
H. Peter Anvin <[EMAIL PROTECTED]> wrote:
>By author: Konrad Mierendorff <[EMAIL PROTECTED]>
>>
>> Joe Pfeiffer wrote:
>> > You have to have LILO. A bit more precisely, you have to have a boot
>> > loader -- something sitting in the boot sector that tells the machine
>> > how to load the OS. The normal boot loader for Linux is LILO.
>>
>> I'm wondering if it's possible to "dd" the kernel-image, like it is
>> possible on floppy disks.
>>
>
>That works only for floppy disks. On anything but floppies you need
>LILO, LOADLIN, or SYSLINUX.
See
http://freshmeat.net/news/1998/07/22/901163113.html
It is a boot block which doesn't need a second-stage loader.
--
Nathan Myers
[EMAIL PROTECTED] http://www.cantrip.org/
------------------------------
From: "BongShin Choi" <[EMAIL PROTECTED]>
Subject: I can't make zImage with "make zImage"
Date: Mon, 09 Aug 1999 04:40:15 GMT
Hi, dear,
After I done "make dep", ran "make zImage as README file included
"linux-2.3.12.tar.gz".
But, when I run "make zImage", got a message with "Error 1", "Error 2".
What's problem?
Who can answer me?
It's a error message followed.
===============================
gcc -D__KERNEL__ -I/usr/src/linux-2.3.12/include -Wall -Wstrict-prototypes -
O2
-fomit-frame-pointer -D__SMP__ -pipe -fno-strength-reduce -m486 -DCPU=486
-c -o hd.o hd.c
hd.c: In function `hd_ioctl':
hd.c:623: `BLKPG' undeclared (first use this function)
hd.c:623: (Each undeclared identifier is reported only once
hd.c:623: for each function it appears in.)
hd.c:624: warning: implicit declaration of function `blk_ioctl'
make[3]: *** [hd.o] Error 1
make[3]: Leaving directory `/usr/src/linux-2.3.12/drivers/block'
make[2]: *** [first_rule] Error 2
make[2]: Leaving directory `/usr/src/linux-2.3.12/drivers/block'
make[1]: *** [_subdir_block] Error 2
make[1]: Leaving directory `/usr/src/linux-2.3.12/drivers'
make: *** [_dir_drivers] Error 2
[root@sc2175 linux]#
[root@sc2175 linux]#
============================
------------------------------
From: "BongShin Choi" <[EMAIL PROTECTED]>
Subject: I can't make zImage.
Date: Mon, 09 Aug 1999 04:44:09 GMT
When I run "make zImage", got a message with "Error 1", "Error 2".
What's problem?
Who can answer me?
It's a error message followed.
===============================
gcc -D__KERNEL__ -I/usr/src/linux-2.3.12/include -Wall -Wstrict-prototypes -
O2
-fomit-frame-pointer -D__SMP__ -pipe -fno-strength-reduce -m486 -DCPU=486
-c -o hd.o hd.c
hd.c: In function `hd_ioctl':
hd.c:623: `BLKPG' undeclared (first use this function)
hd.c:623: (Each undeclared identifier is reported only once
hd.c:623: for each function it appears in.)
hd.c:624: warning: implicit declaration of function `blk_ioctl'
make[3]: *** [hd.o] Error 1
make[3]: Leaving directory `/usr/src/linux-2.3.12/drivers/block'
make[2]: *** [first_rule] Error 2
make[2]: Leaving directory `/usr/src/linux-2.3.12/drivers/block'
make[1]: *** [_subdir_block] Error 2
make[1]: Leaving directory `/usr/src/linux-2.3.12/drivers'
make: *** [_dir_drivers] Error 2
[root@sc2175 linux]#
[root@sc2175 linux]#
============================
------------------------------
From: [EMAIL PROTECTED] (Chris Gregory)
Subject: as86/ld86?
Reply-To: [EMAIL PROTECTED]
Date: Mon, 09 Aug 1999 05:05:28 GMT
I'm building a new linux system from source. I thought I had everything I
need to rebuild the kernel (2.2.10) on my current system, but I don't seem
to have as86 or ld86. I checked the tar of my old system and there was an
as86 but no man page for it. I can't seem to find any source for this.
How do as86 and ld86 differ from as and ld? I have the gnu versions of
these. Where can I get source for current as86 and ld86? Can I emulate the
functionality of these with options to gnu ld and as?
Chris G.
------------------------------
From: Eugene Morozov <[EMAIL PROTECTED]>
Subject: Re: Device driver programming and C++
Date: 09 Aug 1999 09:31:18 +0400
"Andrey Fisunenko" <[EMAIL PROTECTED]> writes:
> >> Where can I read more detail g++ information like that?
> >You'd better read `info gcc'.
> >Eugene
> >
> >--
>
>
> unfortunely `info gcc' contens the same info as 'man gcc'
It means that your gcc was incorrectly installed . Try to find
where gcc installed its info pages and run `info --file=gcc.info'
in this directory.
Eugene
--
Email: <jmv @ lucifer dorms spbu ru> Homepage: http://lucifer.dorms.spbu.ru
------------------------------
Subject: "Domain validation" for reading tape archive???
From: [EMAIL PROTECTED] (Josh Stern)
Date: Mon, 09 Aug 1999 06:11:14 GMT
What is "domain validation" in the context of reading
a tar archive tape and how do I get rid of it?
This question arose for me in the following context:
I'm making some DAT tapes at work of files on a Linux
box, using a tape attached to a SunOS machine. For
directories that are nfs mounted this only involves
normal tar commands. However, for directories that
are not nfs mounted I piped the output of tar
running on the Linux machine to rsh "dd of=/dev/tape"
on the Sun box. My DAT drive at home can read
both tapes (so far so good) however, any and every
operation involving only the tape that was made using rsh
is *slow* as molasses, and when I look in the syslog
I find messages like this:
Aug 10 00:34:32 Atlas kernel: (scsi0:0:5:0) Performing Domain validation.
Aug 10 00:34:32 Atlas kernel: (scsi0:0:5:0) reducing SCSI transfer speed
due to Domain validation failure.
This is only for the rsh tape and it doesn't have anything
obvious to do with security on my machine (e.g. this happens
even if I am just doing tar tvf to list the contents of
the tape). Can anyone explain what is going on here and possibly
suggest a workaround.
Thanks,
- Josh
------------------------------
From: Vladimir <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.hardware,redhat.hardware.arch.intel
Subject: DIGITAL output! Matrox G200.
Date: Mon, 09 Aug 1999 06:30:43 GMT
Hi!
I have Matrox G200, 8MB RAM, AGP with DIGITAL video output only.
(repeat "DIGITAL output"). My display is IBM TFT, model T55D.
I'm in despair. Is there anybody can help me?
Useless:
1. change XF86 server with 3.3.4 (the newest one!)
2. change video card with DIGITAL/ANALOG Matrox (double output).
Problem is not in TFT. It works with new card.
3. change several Matox drivers.
4. change Linux distribution. (RedHat 6.0, SuSE 6.0, Slackware.... .)
Please, help me! I love Linux!
Thanks!
================== Posted via CNET Linux Help ==================
http://www.searchlinux.com
------------------------------
From: [EMAIL PROTECTED] (Villy Kruse)
Crossposted-To: comp.os.linux.misc
Subject: printk ???
Date: 9 Aug 1999 10:43:35 +0200
In article <[EMAIL PROTECTED]>,
David Warren <[EMAIL PROTECTED]> wrote:
>Could this have to do with the "Use versioning on kernel syms" setting?
>Just a wild idea...
>
Try grep printk from /proc/ksyms on a system with versioning enabled
and get something like
001141b0 printk_Rad1148ba
So your wild idea is confirmed.
BTW: it has nothing to do wich c++ mangling.
Villy
------------------------------
From: Etienne Lorrain <[EMAIL PROTECTED]>
Subject: Re: The beginnings of a protected-mode boot loader
Date: Mon, 09 Aug 1999 10:20:19 +0100
"H. Peter Anvin" wrote:
> Konrad Mierendorff <[EMAIL PROTECTED]> wrote:
> > Etienne Lorrain wrote:
> > > It is boring, I do not have extended disk BIOS but have IDE disks
> > > with more than 1024 cylinder, even bigger than 8 Gb...
> >
> > But IMHO your kernel-image has to be below the 1024th cylinder
> >
>
> Yes, if you don't have EBIOS then you need to keep all of /boot below
> cylinder 1024. lbcon will use EBIOS if it is present, but if it
> isn't, there just ain't anything to do.
> -hpa
Hi,
I remember a message you wrote approx one year ago, saying that
to be able to remove this limit, one should write an IDE driver
in the 512 first bytes, and you said it was approximately
impossible, but if someone want to try...
Maybe you should have a look at the file "boot.c" in gujin...
Getting GCC to work for real mode was not the most difficult
part of the "gujin" game!
Note also that the "to be written in last step" installer
could detect 8 Gbytes disk BIOSes, the ones which increases
number of heads or the ones which put the two upper bits of
cylinder in head bit 6 and 7, and the first bootloader
can be initialised with the right data in each case,
but this is only when the IDE driver would not work.
By the way, I installed RedHat 6 this week-end, and after
saying all the bad words because some of the packages in
Eridani distribution were compiled for something bigger than
386 (like RPM.*.rpm) , I noticed "gujin" would not link
because of an (unexplained) inter-segment gap of 0x10 bytes.
Something like (in Makefile, part boot.lnk) fixes it,
but it is dirty (only for RH6 based systems):
- . = $(NbBootSect)*512-SIZEOF(.text); \n\
+ . = $(NbBootSect)*512-SIZEOF(.text)-0x10; \n\
Also EGCS-1.1.2 makes bigger executable than GCC-2.8.1,
even with "-Os" flags, so you have to select less "DEBUG",
and DOSEMU of RH6 seems to behave strangely with gujin.
Have a nice day,
Etienne.
Note: I am sorry to use your thread to make my advertisement,
I hope you will forgive me.
------------------------------
From: "Andrey Fisunenko" <[EMAIL PROTECTED]>
Subject: Re: Device driver programming and C++
Date: Mon, 9 Aug 1999 12:16:42 +0300
>> unfortunely `info gcc' contens the same info as 'man gcc'
>It means that your gcc was incorrectly installed . Try to find
>where gcc installed its info pages and run `info --file=gcc.info'
>in this directory.
>Eugene
I learn and reread info gcc very carefully. And it was installed
successfully. But...
The options
-fno-rtti that was adviced from the newsgroup
and -fno-exceptions (i.e. disbling exception handling) that was
experimentaly found after some attempts.
there are not in info gcc. (I try to grep entry: "rtti" by internal means
of info )
It is a bad practice to hide so valueble options.
Andrey.
------------------------------
From: Christoph Martin <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.hardware
Subject: Re: Toshiba DVD-RAM and Linux
Date: 09 Aug 1999 11:58:14 +0200
Christian Mund <[EMAIL PROTECTED]> writes:
>
> Hello,
> I am the proud owner of a TOSHIBA SD-W1101 DVD-RAM Drive and want to know,
> if there is anyone out there, who is able to use this Drive with Linux.
>
> The Drive works in a 2-LUN Mode, LUN0 = CDROM and LUN1 = Optical Device.
> Reading from a DVD-Media works fine, but when i try to write, the
> Drive hangs up the SCSI-Bus.
> I tried to track this problem a little bit, but i am not so familiar with
> debugging Kernels .
> dmesg says : "aborting command due to timeout : write(6)
> timed out: reset"
> Using the strace-command i can see, that thwe Device-Hang occurs, when the
> Kernel does an fsync.
>
Similar problems here. Linux hangs after writing some data. Same
happens when reading large chunks from DVD-Rom.
Christoph
--
============================================================================
Christoph Martin, Uni-Mainz, Germany
Internet-Mail: [EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Kees J Bot)
Subject: Re: My first linux program: non-bios boot loader
Date: Mon, 9 Aug 1999 11:45:32 +0200
In article <[EMAIL PROTECTED]>,
Neil Koozer <[EMAIL PROTECTED]> wrote:
>
>Hey, this is interesting... LILO (version 20) begins with:
>
>cli
>jmp start
>
>and LILO (version 19) begins with:
>
>jmp start
>
>A comment on the 'cli' line say "NT blows up without this"
Hmm, that would make one think that a CLI instruction is the necessary
fix, while in reality the first instruction just has to be a non-jump.
On my tests I even went so far to use DOS 'fdisk /mbr' to install
ordinary MBR code, then I changed the first byte into a jump
instruction. When NT was booted (using a floppy rigged to boot a disk
partition) then it would crash. Without the first byte changed it would
work. In both cases the MBR code wasn't even executed, so it's just the
contents that made NT puke, not the code. I even filled the MBR with
446 zeros: NT boots. Change first byte to a jump: NT crashes.
What I haven't figured out yet is what NT changes on the MBR if you run
its disk manager utility. There seem to be some bytes changed just
above the partition table. One of my replacement master bootstraps is
436 bytes, so I hope NT doesn't need more than 10.
--
Kees J. Bot, Systems Programmer, Sciences dept., Vrije Universiteit Amsterdam
Minix: http://www.cs.vu.nl/pub/minix/ ftp://ftp.cs.vu.nl/pub/minix/
Minix-vmd: http://www.Minix-vmd.org/ ftp://ftp.Minix-vmd.org/
------------------------------
From: Etienne Lorrain <[EMAIL PROTECTED]>
Subject: Re: My first linux program: non-bios boot loader
Date: Mon, 09 Aug 1999 11:05:39 +0100
"H. Peter Anvin" wrote:
>
> Followup to: <[EMAIL PROTECTED]>
> By author: Neil Koozer <[EMAIL PROTECTED]>
> In newsgroup: comp.os.linux.development.system
> >
> > Charles Sullivan wrote:
> >
> > > Sounds like a great idea. I assume it is consistant with
> > > having other operating systems on the system.
> >
> > I think so, but then I don't know why NT can't handle lilo being in the
> > MBR. I would assume that nuni would work in /hda1 (for example) the same as
> > lilo for NT users.
> >
>
> Because NT stores nonstandard data in the MBR (quite possibly to
> intentionally screw up other OSes).
>
> -hpa
> --
> <[EMAIL PROTECTED]> at work, <[EMAIL PROTECTED]> in private!
Not only that, there is a better joke:
Some of the SCSI disk BIOSes, like this well known company
which claims to support LINUX, is playing with the bootsector
BEFORE jumping to 0x7c00. At initialisation time, it rewrite
few bytes in the bootsector, they are usually letters like
'C', 'D', 'E' - probably to inform Window NT the order of
the drives. You probably can get the specification by getting
a NDA with micro$oft.
How to check it:
- get a Linux bootable floppy which works
- boot linux, and clear with "dd /dev/zero" the first
512 bytes of the booting disk, i.e. C:
- shutdown and reboot Linux with the floppy, do not
try to boot anything in between.
- read the first 512 bytes of the "new" bootsector,
and have fun.
I begin to understand why a new software is downloaded
to the SCSI card by the Linux driver at boot...
To be fair, the address rewritten were reserved by
MS-DOS and LILO is using them for another purpose,
to store its own code/data.
By the way, IHMO a bootloader should be able to
handle BIOS, EBIOS and IDE interface, and do a checksum
of its code to not be modified by a virus, whatever the
name of the virus.
Etienne.
------------------------------
From: [EMAIL PROTECTED] (H. Peter Anvin)
Subject: Re: My first linux program: non-bios boot loader
Date: 9 Aug 1999 09:37:25 GMT
Reply-To: [EMAIL PROTECTED] (H. Peter Anvin)
Followup to: <[EMAIL PROTECTED]>
By author: Neil Koozer <[EMAIL PROTECTED]>
In newsgroup: comp.os.linux.development.system
>
> Hey, this is interesting... LILO (version 20) begins with:
>
> cli
> jmp start
>
> and LILO (version 19) begins with:
>
> jmp start
>
> A comment on the 'cli' line say "NT blows up without this"
> This would explain how someone recently reported that he put LILO in the MBR
> without affecting NT.
>
Useful thing to know...
-hpa
--
<[EMAIL PROTECTED]> at work, <[EMAIL PROTECTED]> in private!
------------------------------
From: [EMAIL PROTECTED] (Martin Boening)
Subject: Re: bzip2 compressed kernel
Date: 9 Aug 1999 10:27:11 GMT
Hi there,
[EMAIL PROTECTED] (H. Peter Anvin) writes:
>Followup to: <7obnmg$912$[EMAIL PROTECTED]>
>By author: [EMAIL PROTECTED] (Martin Boening)
>In newsgroup: comp.os.linux.development.system
>>
[ my blabbering deleted ]
>bzImage has nothing to do with bzip. b = big.
AhAAah! Thanks for clearing up this misconception of mine. Since this
is so, and obviously something has been done to work around the 640K
limit before the kernel itself is running - why not do it always and
dispense with zlilo/zimiage/zdisk and friends?
Just wondering...
So long,
Martin
--
Martin Boening, MB3792 | EMail: [EMAIL PROTECTED]
SBS SCN VAS, Otto-Hahn-Ring 6, D-81739 Muenchen (Perlach), Germany
Phone: +49 896 364 2904 FAX: +49 893 365 1031
------------------------------
** 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
******************************
