Linux-Development-Sys Digest #934, Volume #6 Mon, 5 Jul 99 22:13:55 EDT
Contents:
Problems reading whole CD with SCSI-CDROMs (Sven Anders)
LCD (Christelle Van-Den-Broucke)
Re: help:why program can be run inside gdb but can not in command line (Bernd
Strieder)
Re: usleep() (Daniel R. Grayson)
Re: Need for a 128-bit (Unix or other) OS? (susano)
Move command source ("Steve B.")
Re: binutils compile error (Ulrich Drepper)
Re: binutils compile error (Eric Buddington)
Re: Need for a 128-bit (Unix or other) OS? (Johan Kullstam)
serial driver problem (Kuang-chun Cheng)
CD-ROM File Time Bug (Terry Boldt)
Re: Need for a 128-bit (Unix or other) OS? ("Kelly_Robinson")
Device Help ("Dmitriy Bulgakov")
Re: Why we are still holding on to X Windows (Mike McDonald)
Re: Move command source (John McKown)
Re: driver for AMCC S5933 (Eric Hegstrom)
Testing cvs-versions of glibc ([EMAIL PROTECTED])
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (Sven Anders)
Crossposted-To:
comp.os.linux.misc,de.comp.os.linux.misc,de.comp.os.linux.hardware,linux.act.scsi,linux.dev.scsi,list.linux-activists.scsi
Subject: Problems reading whole CD with SCSI-CDROMs
Date: 4 Jul 99 22:00:38 GMT
When I try to read the iso image of a whole CD with 'dd' (something like:
dd if=/dev/sr0 of=iso.img count=`isosize /dev/scd0`) I didn't get the
whole image. They are always incomplete and I got errors, if I tried to access
the last files on the image.
I tried it with my SCSI Toshiba CDROM and SCSI Phillips CD-Writer.
It works perfectly with my ATAPI CDROM. I use a NCR810 SCSI Controller.
Is it an error of the SCSI drivers or do I something wrong ?!
Please send me the answer a mail too.
My e-mail address is: [EMAIL PROTECTED]
Thanks in advance,
Regards
Sven
--
>>> "I am the embodiment of modern medicine." <<<<<<< Der Dativ ist dem <<<<<<<
>>>>>> The holographic doctor, USS-Voyager <<<<<<<<<< Genitiv sein Tod. <<<<<<<
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>><<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
------------------------------
From: Christelle Van-Den-Broucke <[EMAIL PROTECTED]>
Subject: LCD
Date: Mon, 05 Jul 1999 10:47:55 +0200
Does anyone use a LCD optrex DMC 40218?
I'd like to know how to connect it on the parallel port.
------------------------------
From: Bernd Strieder <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.development.apps
Subject: Re: help:why program can be run inside gdb but can not in command line
Date: Mon, 05 Jul 1999 09:49:35 +0200
Hello,
y chen wrote:
>
> Hello , there ,
> I had a program and i can run it inside gdb
> without any problem.
> However, When i run it in command line,
> it seg fault.
> Does anyone have same experience and
> tell me how to fix it?
> Thanks!
FAQ?
this is most common for C programmers, and others, too There could be
all kinds of errors causing this behaviour. I remember that I had double
free(), free()'s with a following access to the freed memory, forgotten
initializations of pointers amongst other errors causing such behaviour.
The occurrence of seg-faults depends on the input data and even on the
machine you are running on. This is because one does things that lead to
undefined behaviour, as defined by the programming language. Segfaults
are undefined behaviour of your app catched by the OS to avoid damages
to others.
I would recommend you to use some malloc-debugging lib, like
ElectricFence (libefence), or dmalloc. I don't know URLs, but the search
engines are your friend. These libs indicate various malloc/free related
problems.
One strategy I often did, was to overwrite the block to be freed with
garbage 0xaa's. So the segfault happens earlier if there are pointers in
this block. It is easier to track it down.
And test with many different inputs, perhaps generated by random. If
you have seg-faulting input data try to find the point where it crashes
in the core dump. Then trace in the debugger from the beginning. gdb
itself might be too crude, so use a graphical debugger like ddd, ups or
at least xxgdb.
HTH
Bernd Strieder
------------------------------
From: [EMAIL PROTECTED] (Daniel R. Grayson)
Subject: Re: usleep()
Date: 05 Jul 1999 07:43:53 -0500
Peter Pointner <[EMAIL PROTECTED]> writes:
> Daniel R. Grayson <[EMAIL PROTECTED]> wrote:
>
> [snip]
>
> > Thanks for the precise explanation about how usleep() works.
>
> > Have you ever wished usleep() were accurate? It's possible, because select()
> > with timeout is accurate.
>
> Yes, I wished. But I doubt select() is better. Did you test that and how
> accurate is it? The reason I doubt it is that I can't see how it might be
> done.
>
> Peter
This is a good question, so I tested it more carefully, and it turns out that
select has an overhead of 10 ms and usleep has an overhead of 20 ms, both of
which are explained by counting clock ticks. (So select wasn't so accurate
after all, sorry!)
=============================================================================
With select:
% gcc -O foo.c
% time a.out
real 0m5.015s
user 0m0.000s
sys 0m0.000s
% cat foo.c
#include <sys/times.h>
#include <ctype.h>
#include <signal.h>
#include <stdio.h>
#include <stddef.h>
#include <unistd.h>
#include <curses.h>
#include <sys/time.h>
#include <sys/types.h>
#define udelay 1
main() {
int fd[2];
int i;
pipe(fd);
for (i=0; i<500; i++) {
fd_set fds;
struct timeval timeout = {0, udelay};
FD_ZERO(&fds);
FD_SET(fd[0],&fds);
select(fd[0]+1, &fds, 0, 0, &timeout);
}
}
With usleep():
% gcc -O bar.c
% time a.out
real 0m10.011s
user 0m0.010s
sys 0m0.000s
% cat bar.c
#include <sys/times.h>
#include <ctype.h>
#include <signal.h>
#include <stdio.h>
#include <stddef.h>
#include <unistd.h>
#include <curses.h>
#include <sys/time.h>
#include <sys/types.h>
#define udelay 1
main() {
int i;
for (i=0; i<500; i++) {
usleep(udelay);
}
}
------------------------------
From: [EMAIL PROTECTED] (susano)
Crossposted-To: alt.os.linux,comp.os.linux.advocacy,comp.unix.advocacy
Subject: Re: Need for a 128-bit (Unix or other) OS?
Date: Mon, 05 Jul 1999 18:32:00 GMT
It's even worse than that, it means the rest of the Intel programmers
catching up with th capabilities of the 80286.
32-operating sysems are differentiated from 16-bit OSes by their
support for 3 things: multitasking, protected memory spaces, and
virtual memory. All things that the 80286 provided support for,
>The phrase "32-bit operating system for intel processors" just means it
>takes advantage of intel finally catching up with the rest of the
>computing world.
------------------------------
From: "Steve B." <[EMAIL PROTECTED]>
Subject: Move command source
Date: Mon, 5 Jul 1999 18:35:30 -0700
Where can I find the source to basic commands like mv. I have RH 6.0 and
loaded the source, but not finding the source to commands. The reason I
am looking is a bet with someone whether mv does a copy and delete or just
adjusts the inodes. In early days of Unix all moves were actually
copy/delete, I say that is only done today when crossing partitions or
disks. Like to find the source to see who's right.
Steve B.
------------------------------
From: Ulrich Drepper <[EMAIL PROTECTED]>
Subject: Re: binutils compile error
Date: 05 Jul 1999 12:08:21 -0700
Reply-To: [EMAIL PROTECTED] (Ulrich Drepper)
[EMAIL PROTECTED] (Eric Buddington) writes:
> Undefined __register_frame_info rings a bell as an annoying
> glibc-2/2.1 incompatibility.
This has nothing to do with incompatibilities in glibc. It's a
compiler issue. If you want to blame people make damn sure you know
what you are talking about.
--
===============. drepper at gnu.org ,=. 1325 Chesapeake Terrace
Ulrich Drepper \ ,-------------------' \ Sunnyvale, CA 94089 USA
Cygnus Solutions `--' drepper at cygnus.com `------------------------
------------------------------
From: [EMAIL PROTECTED] (Eric Buddington)
Subject: Re: binutils compile error
Reply-To: [EMAIL PROTECTED]
Date: 5 Jul 1999 14:36:44 -0500
In article <7lo0k3$qqm$[EMAIL PROTECTED]>, Klaus Schneider wrote:
>Vic Mulyk ([EMAIL PROTECTED]) wrote:
>> Klaus Schneider wrote:
>> >
>> > Hi all,
>> >
>> > trying to compile and install binutils 2.9.1
>> > the compilation with �make� went ok, but when
>> > �make install� I got a strange error:
>> >
>> > ranlib: error in loading shared libraries
>> > : undefined symbol: __register_frame_info
Klaus,
Undefined __register_frame_info rings a bell as an annoying
glibc-2/2.1 incompatibility. And I think there are a few other
now-undefined symbols as well. I don't know why they can't just throw
in a dummy symbol to avoid the error, or bump of the major version
again, since that's what major versions are for...
HTH,
Eric
------------------------------
Crossposted-To: alt.os.linux,comp.os.linux.advocacy,comp.unix.advocacy
Subject: Re: Need for a 128-bit (Unix or other) OS?
From: Johan Kullstam <[EMAIL PROTECTED]>
Date: 05 Jul 1999 14:37:06 -0400
[EMAIL PROTECTED] (susano) writes:
> It's even worse than that, it means the rest of the Intel programmers
> catching up with th capabilities of the 80286.
>
> 32-operating sysems are differentiated from 16-bit OSes by their
> support for 3 things: multitasking, protected memory spaces, and
> virtual memory. All things that the 80286 provided support for,
that has nothing to do with 32 bitness.
> >The phrase "32-bit operating system for intel processors" just means it
> >takes advantage of intel finally catching up with the rest of the
> >computing world.
intel still hasn't caught up. however this is more microsoft's fault
than intel's. the i386 was introduced in the late '80s but it took
microsoft until 95 to come up much for the 32 bit mode. most of the
world has moved to 64 bit arches. also nearly everyone else uses a
risc type cpu. not so with intel.
--
J o h a n K u l l s t a m
[[EMAIL PROTECTED]]
Don't Fear the Penguin!
------------------------------
From: Kuang-chun Cheng <[EMAIL PROTECTED]>
Subject: serial driver problem
Date: 30 Jun 1999 08:10:45 GMT
Hi,
I have a GaussMeter (FWBell 9200) which support serial port
output. The serial port configuration of GaussMeter is
B1200
8 Bit
1 Stop
No Parity
The meter act as a talker which usually connect to a terminal or
printer. The connection can be achieve by only two wire (ground and TD).
I connect the meter to Win95/98/NT's HyperTerminal, it works OK. The
HyperTerminal keep printing the output value of GaussMeter. However, when
I use
cu -1200 -l /dev/ttyS1
on Linux. Nothing happen :(
I know the serial driver of Linux is stable. But some how when I
try to control some instruments using serial port under Linux, I always have
problems. For example, the tcdrain() seems not work too. And also, I can't
find any method to set/clear RTS line alone ...
Does serial driver of Linux really OK ?? Or does someone has
similar problems ... Thanks.
Kuang-chun Cheng
[EMAIL PROTECTED]
------------------------------
Date: Mon, 05 Jul 1999 19:18:00 -0400
From: Terry Boldt <[EMAIL PROTECTED]>
Reply-To: [EMAIL PROTECTED]
Subject: CD-ROM File Time Bug
I have recently encoutered a problem with listing file times on the
CD-ROM. I have
cross checked the file times between MS-DOS 6.2, OS/2 and Linux. MS-DOS
and OS/2
both agree on the time a file was written to the CD-ROM, but Linux adds
3 hours
to the time. For example under MS-DOS and OS/2 the file time is listed
as 1:15 pm.
Using the command 'ls -l --full-time /mnt/cdrom' under Linux the file
time is listed
as 16:15. Since both MS-DOS and OS/2 (using both the 'dir' and the GNU
'ls -l --full-time' commands) both agree on the file time, I assume that
Linux as the
odd man out is incorrect. I have checked the file times listed under
Linux for files
on the hard disk and floppies, they are correct for when I created the
files. So I assume
that the problem is in the CD-ROM driver.
I am using Red Hat 5.2 and the CD-ROM is an IDE ATAPI drive.
Is this a known bug and if so has the problem been corrected?
Getting file times that agree across multiple platforms is crucial to my
application
and so, of course, this problem came to my attention very quickly. I
don't need
a correction right now, but must know if it is known and corrected or
being corrected. If the problem is not known, whom would I alert to the
problem. I assume that others would be interested in correct file times
also.
Thank you for any input you may give.
------------------------------
From: "Kelly_Robinson" <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux,comp.os.linux.advocacy,comp.unix.advocacy
Subject: Re: Need for a 128-bit (Unix or other) OS?
Date: Mon, 5 Jul 1999 16:54:58 -0500
susano wrote in message <3783f92f.16820340@news-server>...
>It's even worse than that, it means the rest of the Intel programmers
>catching up with th capabilities of the 80286.
>
>32-operating sysems are differentiated from 16-bit OSes by their
>support for 3 things: multitasking, protected memory spaces, and
>virtual memory. All things that the 80286 provided support for,
>
>>The phrase "32-bit operating system for intel processors" just means it
>>takes advantage of intel finally catching up with the rest of the
>>computing world.
>
Huh? Intel has always been behind Motorola in terms of technology, but it
is the responsibility of the OS maker (or the computer maker if they were
responsible enough to do the entire computer itself) to make up-to-date
operating systems.
------------------------------
From: "Dmitriy Bulgakov" <[EMAIL PROTECTED]>
Subject: Device Help
Date: Mon, 5 Jul 1999 09:09:33 +0300
What debugger can I use for debugging a device driver?
please help me.
------------------------------
From: [EMAIL PROTECTED] (Mike McDonald)
Crossposted-To: comp.os.linux.development.apps
Subject: Re: Why we are still holding on to X Windows
Date: 6 Jul 1999 00:56:58 GMT
Reply-To: [EMAIL PROTECTED]
In article <[EMAIL PROTECTED]>,
Mario Klebsch <[EMAIL PROTECTED]> writes:
> .. and I have read it several years ago, as I have done with the open
> look styleguide, too. I found lots of their principles in MacOS and in
> Windows 95, but a lot of X11 programs still do ignore these guides
> completely. :-(
>
> And the developer community would have to agree on one of them. And year
> after year lots of programs are written, each only acording to the autors
> stype giude. Delaying only makes the problem bigger and bigger.
Do you know why there are so many different styles? It isn't always just
because the programmers were too lazy to follow one of the existing guides.
Quite often, it is because the existing guides and toolkits are inadequate,
for a number of reasons. If you want someone to dictate the "one true GUI",
use Windows. I personally don't think any of the GUI tolkits have gotten to a
point where they solve a majority of the UI issues, let alone enough to be
cast in stone.
Mike McDonald
[EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (John McKown)
Subject: Re: Move command source
Date: 6 Jul 1999 00:38:00 GMT
On Mon, 5 Jul 1999 18:35:30 -0700, Steve B. <[EMAIL PROTECTED]> wrote:
>Where can I find the source to basic commands like mv. I have RH 6.0 and
>loaded the source, but not finding the source to commands. The reason I
>am looking is a bet with someone whether mv does a copy and delete or just
>adjusts the inodes. In early days of Unix all moves were actually
>copy/delete, I say that is only done today when crossing partitions or
>disks. Like to find the source to see who's right.
>
>Steve B.
The source should be on disk #2 in file fileutils-4.0-1.src.rpm. You can
look at the source if you want. It seems a bit convoluted to me. I did
look at it. I couldn't be sure, so I cheated (of course). I used the
command:
strace -o trace mv file subdirectory/
This moves "file" from the current directory to "subdirectory". When
I looked at the trace, there were a few entries, but towards the end
was the entry:
rename("file","subdirectory/file")
rename is a system function which moves directory entries. It does not
appear to actually do an I/O other that to the directory.
------------------------------
From: Eric Hegstrom <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.linux.hardware,comp.os.linux.setup,comp.os.linux.misc,comp.os.linux.development.apps
Subject: Re: driver for AMCC S5933
Date: Mon, 05 Jul 1999 17:55:18 -0700
Yes,
A nice gentleman from Italy shared some development code he had worked
on for the s5933. It seems to work pretty darn well.
His name is Andrea Cisternino <[EMAIL PROTECTED]>
His source code was at
http://pcape1.pi.infn.it/~acister/dev/driver.html
If you have trouble I can probably find a copy to send you.
Peace,
Eric
"Dr. Oleg P. Nikolayev" wrote:
>
> Dears:
>
> Is there driver for AMCC S5933 (Linux Red Hat for PC Pentium)?
> If there is, as it to teceive?
>
> Thanks, Oleg
>
> --
> Dr. Oleg P. Nikolayev
> Graduate School of Applied Science - Voice: 972-2-65-84694
> Dept. of Psychology - Voice: 972-2-58-81089, Fax: 972-2-5825659
> The Hebrew University of Jerusalem - Email: [EMAIL PROTECTED]
--
Eric Hegstrom .~.
Senior Software Engineer /V\
Sonoran Scanners, Inc. // \\ L I N U X
[EMAIL PROTECTED] /( )\ >don't fear the penguin<
520-617-0072 x402 ^^-^^
------------------------------
From: [EMAIL PROTECTED]
Subject: Testing cvs-versions of glibc
Date: Tue, 06 Jul 1999 00:19:20 GMT
Hi,
Yesterday I decided to try out glibc from the cvs-repository. It
compiled without hassle, and also 'make check' worked. Then I tried to
compile some test applications. But the binaries I produce from gcc
tells me that it can't find the file when trying to run it. (That is
also what ldd reports)
Is there a FAQ or a simple explanation of how to set up a test-version
of glibc? The glibc-FAQ was helpful but didn't get me all the way to a
runnable executable.
A short description of what I've done:
compiled glibc and installed it in a separate cvs-glibc directory.
export LD_LIBRARY_PATH=${HOME}/cvs-glibc/lib
export PATH=${HOME}/cvs-glibc/bin:${PATH}
gcc -Wl,--dynamic-linker=${HOME}/cvs-glibc/lib/ld-linux.so.2
-I${HOME}/cvs-glibc/include hello.c -L${HOME}/cvs-glibc/lib
I remember that it wasn't so hard compiling libc5 and libc6 applications
on the same system so I must be doing something really stupid. :-/
/Jonas U
Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
------------------------------
** 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
******************************
