Linux-Development-Sys Digest #326, Volume #6 Sun, 24 Jan 99 10:19:19 EST
Contents:
Re: Lightnux, the slim Linux (Christopher B. Browne)
Re: 2.2pre9 won't compile (Stephen Sill)
Re: Direct Sound? (Stephen Sill)
Re: glibc2.0.7 ldd {somefile} FAILS (Pinwu Xu)
Re: 2.2pre9 won't compile (Peter S. Fales)
Problem memory mapping many files (No free inodes) (Andrew Pipkin)
Q: Debugging Linux Device Drivers ("Bertrand Lee")
Re: Lightnux, the slim Linux (Douglas Kilpatrick)
Re: BeOS and Linux (Nathan Myers)
Re: BeOS and Linux ("Axel Winter")
Re: S3 ViRGE + KDE ([EMAIL PROTECTED])
numeric keypad and rxvt (Franck ZOCCOLO)
Re: Why I'm dumping Linux, going back to Windblows (Todd Ostermeier)
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (Christopher B. Browne)
Subject: Re: Lightnux, the slim Linux
Reply-To: [EMAIL PROTECTED]
Date: Sun, 24 Jan 1999 03:01:23 GMT
On 24 Jan 1999 00:14:27 GMT, Pascal Ferrari <[EMAIL PROTECTED]> posted:
>Yesterday night I had a dream :
>
>I was visiting one of my clients' office and was looking at the people
>working there :
>- 8 secretaries typing mail as usual on their old x386 Win3.1 machine,
>- 2 accountants entering book items from their term emulator using my
>remote char-based accounting application running on the Linux Server,
>- 3 managers using their wonderful Pentium Pro box under Win98 one hour per
>day for just playing Doom.
>
>In my dream I was unplugging the x386 machines, leaving to the users their
>monitor, keyboard and mouse and plugging these peripherals back on a small
>and flat box I placed under their monitor. This box was just offering a
>VGA, a keyboard, a mouse and a network connection plugs : no floppy drive,
>no CD and even no hard drive.
All pretty possible right now.
The unfortunate thing is that there isn't any reasonable alternative
at this time to disk for "fairly massive" amounts of data. There was
once a theory that "bubble memory," a sort of serial-bus-based memory
system, would replace disk. Real RAM got cheap fast, and you can now
buy bubble memory in small quantities for high prices for any "embedded
applications" that forcibly need it. (A company bought out the inventory,
and basically sells the existing stock.)
It would be really nice to have something that could hold 100MB of data
in relatively nonvolatile form without the need for any mechanical parts,
with compact size. That's enough for a pretty complete Linux install,
and would be *very* nice for laptops/palmtops/"network computers."
>This box contained its own OS stored in a simple PROM. This OS was a tiny
>Linux Kernel with just the minimal functions for good networking : Web
>browser, terminal emulation and GUI.
Unfortunately, the only reasonable implementation is for this PROM to be
a filesystem with around 16MB of space.
One approach would be to have a battery-backed memory bank that would have
a chunk defined as a RAMdisk with a root filesystem. If it can survive
4 days without power without having to "search the net for a kernel,"
that's probably good enough.
>Not only I could sell these nice boxes at an interesting price, but my
>client gave me for free all their x386 CPUs, thanking me for taking off
>this obsolete stuff. And with this old material I could build new boxes by
>using the inside components, and sell back these boxes in a new client
>office.
They're not going to be happy getting old-looking boxes. And you'll have
a hard time making 'em reliable.
It is probably going to be cheaper to dumpster the old boxes, and build
the new boxes from scratch. A custom motherboard and case will be the
tough part; the point being to make both compact, and integrate NIC
and video onto mobo. In quantity, something with 64MB RAM, rechargable
battery, Tulip, S3V, and low end AMD/Cyrix should be able to be priced
at around $300 or $400.
>But for a workstation, it becomes too much. So, why not to write a very
>light and packaged version of the current Linux, with only the minimal
>functions for networking easily? This "Lightnux" should be no more than 4
>MB if we want to store it on a simple PROM.
I'd rather go with "battery-backed RAM," as that should cost on on the
order of $50, and not require much software revision. Writing new
software is costly.
--
Those who do not understand Unix are condemned to reinvent it, poorly.
-- Henry Spencer <http://www.hex.net/~cbbrowne/lsf.html>
[EMAIL PROTECTED] - "What have you contributed to Linux today?..."
------------------------------
From: Stephen Sill <[EMAIL PROTECTED]>
Subject: Re: 2.2pre9 won't compile
Date: Sun, 24 Jan 1999 03:13:01 GMT
Bill Anderson wrote:
>
> [EMAIL PROTECTED] wrote:
> >
> > I tried compiling 2.2pre9, this is what I got:
> >
> > [ivo@one linux]$ make bzImage
> ...
>
> > /usr/src/kernel/linux/include/asm/smp.h:209: warning: implicit declaration of
>function `GET_APIC_ID'
> > /usr/src/kernel/linux/include/asm/smp.h:209: `APIC_BASE' undeclared (first use
>this function)
> > /usr/src/kernel/linux/include/asm/smp.h:209: (Each undeclared identifier is
>reported only once
> > /usr/src/kernel/linux/include/asm/smp.h:209: for each function it appears in.)
> > /usr/src/kernel/linux/include/asm/smp.h:209: `APIC_ID' undeclared (first use this
>function)
> > make: *** [init/main.o] Error 1
<snip>
I have an ASUS P2B-DS with 2 PII 350's and 256 megs of memory.
2.2.0-pre9 compiled fine, no bugs whatsoever. I patched the 2.2.0-pre8
source tree to 2.2.0-pre9, this could have made a difference.
eitherway, it compiled, and it's FAST.
Stephen
--
Meester, do you vant to buy a duck?
-
Troglodytism does not necessarily imply a low cultural level.
------------------------------
From: Stephen Sill <[EMAIL PROTECTED]>
Subject: Re: Direct Sound?
Date: Sun, 24 Jan 1999 03:19:46 GMT
Dr. Unk wrote:
>
> I was wondering if there was any such drivers for sound, like the ones
> used in Windows. I would like to be able to play multiple sounds with
> out getting an error like, "Device or Resource Busy." Is there any
> place I can go to get these drivers, or get pointers in making them?
>
> Thanks.
My understanding is that DirectSound is a Microsoft API, so it can't be
put in the kernel. But, OSS does have a driver that allows wave mixing
like you are talking about.
www.opensound.com
check it out.
STeve
--
Meester, do you vant to buy a duck?
-
Troglodytism does not necessarily imply a low cultural level.
------------------------------
From: Pinwu Xu <[EMAIL PROTECTED]>
Subject: Re: glibc2.0.7 ldd {somefile} FAILS
Date: Sat, 23 Jan 1999 21:19:08 -0500
Mark Swanson wrote:
>
> ldd always fails. If I:
>
> [root@linux:lib] file /bin/vim
> /bin/vim: ELF 32-bit LSB executable, Intel 80386, version 1, dynamically
> linked, stripped
>
> *this shows it is a dynamically linked executable*
>
> If I then:
>
> [root@linux:lib] ldd /bin/vim
> not a dynamic executable
>
It was stated in some README/INSTALL/FAQ, that you should use the ldd
program from your original system -- if you upgraded from libc5.
------------------------------
From: [EMAIL PROTECTED] (Peter S. Fales)
Subject: Re: 2.2pre9 won't compile
Date: 24 Jan 1999 03:26:59 GMT
Reply-To: [EMAIL PROTECTED]
In article <[EMAIL PROTECTED]>,
Bill Anderson <[EMAIL PROTECTED]> wrote:
>[EMAIL PROTECTED] wrote:
>>
>> I tried compiling 2.2pre9, this is what I got:
>>
>> /usr/src/kernel/linux/include/asm/smp.h:209: `APIC_BASE' undeclared (first use this
>function)
>> /usr/src/kernel/linux/include/asm/smp.h:209: `APIC_ID' undeclared (first use this
>function)
I got the same results, so I tried compiling again with SMP turned
off. That didn't build either (for a different reason) so I switch SMP
back on, and it compiled without a hitch. Don't ask me why, and YMMV.
--
Peter Fales Lucent Technologies, Room 9A-213
N9IYJ 2000 N Naperville Rd PO Box 3033
internet: [EMAIL PROTECTED] Naperville, IL 60566-7033
Remove the "1" to reply work: (630) 979-8031
------------------------------
From: [EMAIL PROTECTED] (Andrew Pipkin)
Subject: Problem memory mapping many files (No free inodes)
Date: 21 Jan 1999 05:21:27 GMT
Reply-To: [EMAIL PROTECTED]
A system lockup can result from using mmap to map pages from a large number
(over 3500) of files. The following program maps the first page of all the
filenames received from standard input. Invoking the program with more than
3500 filenames (this can be achieved by running "find / | <program name>")
will generate the message "VFS: No free inodes - contact Linus" and the
program will hang up. No other programs will be able to access a disk until
the program is killed. If the program is not terminated immediately after the
message is displayed, the system can lose the ability to do any disk I/O,
forcing a reboot. I am running kernel 2.0.36 from RedHat 5.2.
Andrew Pipkin
/* Test program for mmap. Memory map the first page for all filenames
* received via standard input */
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include <fcntl.h>
#include <sys/mman.h>
int main(void)
{
size_t page_len = getpagesize();
char fname[1024];
while(fgets(fname, 1024, stdin))
{ int fdesc;
char *nl = strchr(fname, '\n'); /* Remove trailing \n from filename*/
if(nl)
*nl = '\0';
if((fdesc = open(fname, O_RDONLY)) != -1)
{ char *page = mmap(NULL, page_len, PROT_READ, MAP_SHARED, fdesc,0);
close(fdesc);
}
}
return EXIT_SUCCESS;
}
------------------------------
From: "Bertrand Lee" <[EMAIL PROTECTED]>
Subject: Q: Debugging Linux Device Drivers
Date: Sun, 24 Jan 1999 13:45:03 +0800
Does anyone know of any way to debug Linux device drivers (either in module
or kernel form) by stepping through the code?
I'm looking for something that will let me do what NuMega SoftIce(TM) does
for Windows, ie. step through the code, examine variables, memory, etc.
I'm aware that gdb has some of this capability through remote debugging, but
only for non-Intel platforms, which I am stuck with! So, right now, the main
thing I am doing is printing debug messages. If there is no other way I may
have to write a user-space app to do the debugging through ioctls.
Really appreciate any info. Please include an email reply if possible.
Bertrand Lee.
------------------------------
From: Douglas Kilpatrick <[EMAIL PROTECTED]>
Subject: Re: Lightnux, the slim Linux
Date: Sun, 24 Jan 1999 01:07:29 -0500
Pascal Ferrari wrote:
> This box contained its own OS stored in a simple PROM. This OS was a tiny
> Linux Kernel with just the minimal functions for good networking : Web
> browser, terminal emulation and GUI. When the users plugged on their new
> machine, there was no OS loading delay and no noise. They immediatly got
[snip]
Check out the "Diskless" mini-howto. Its a little out of date; I don't
think the kernel and ramdisk will fit on a single floppy anymore. The
basic idea should still be sound.
Doug
-- [EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Nathan Myers)
Subject: Re: BeOS and Linux
Date: 23 Jan 1999 22:04:01 -0800
John A. Crow<[EMAIL PROTECTED]> wrote:
>I am wondering if anyone has given any thought to the possibility of
>some sort of coordination of development between Linux and BeOS.
>
>My guess would be that the folks at Be (http://www.be.com) have
>given some thought to this, but I never have seen any details.
Yes, they have. They were asked if they would like to fund
development of (GPL'd) drivers that would be portable between
Linux and BeOS, and they declined despite the obvious mutual
benefits.
They are well-funded, though, so they might be persuaded to do
_something_ useful with the money while it lasts. They did toss
some money to Cygnus, some of which has no doubt found its way
into improvements in the compiler/linker/debugger tools we all use.
--
Nathan Myers
[EMAIL PROTECTED] http://www.cantrip.org/
------------------------------
From: "Axel Winter" <[EMAIL PROTECTED]>
Subject: Re: BeOS and Linux
Date: Sun, 24 Jan 1999 13:55:27 +0700
Guys,
I think BeOS is there because Media+Graphic People have very different
requirements then server/database. BeOS has the advantage of offering a more
Plug-n-Play oriented approach towards, installing it and related software. I
don't think a designers are really into installing Linux, X, etc.
However, BeOS greatly increased performence on graphic processing
(rendering, etc.) which normal implemented hw/sw is not capable of ... You
may want to think about BeOS as a better MacOS ... or even a specialized
NeXT Step. This does not mean, you can't do that with Linux - also I don't
think the required graphic tools are available - but I just worry about BeOS
target clients ability to install Linux...
Synergies are: Someone could offer a better integration of BeOS client
services, into Linux Server services. Not unlike this has been done for
Netware and Windows Networking. This would definitely, help BeOS users to
avoide Windows alltogether...
The latter part is something, which could really frustrate MS! If people
would use MacOS or BeOS as client and Linux as Server, there is no part of
MS left in the chain.
Axel
Martin Neumann <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
>"John A. Crow" <[EMAIL PROTECTED]> writes:
>
>> Thus my ill-posed question: *If* BeOS is reasonable for desktop
>> use, and Linux is superb for servers (and many argue the desktop
>> too!) does it make sense to look for a little synergy and
>> cooperation between the two?
>
>Where is the profit for Linux there? As you said, the only thing
>being perhaps better in BeOS than in Linux is the GUI.
>
>I don't think the BeOS GUI is better. The underlying technology of
>X11 is rather good (remote displays etc.). The real desktop is
>perhaps not that sophisticated comparing BeOS and typical X11
>installation with only a window manager and no complete desktop
>environment. But there are already exciting desktop environments
>on the way to everybodys desktops - they're GNOME and KDE.
>
>PS: Having the choice I would always pick GNOME and not KDE, but YMMV
>--
>Martin Neumann \\ [EMAIL PROTECTED]
>EPOS Development \\ 76 B3 C8 55 61 46 34 00 86 87 72 6E C9 8F 38 6D
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: S3 ViRGE + KDE
Date: Sun, 24 Jan 1999 07:54:40 GMT
BLESS YOU! This worked like a charm for me. I thought that I was going to
have to toss KDE because of this. KDE has posted a fix on this, but when I
tried it, it fixed the font problem, but the text showed up colored
strangely...as if it was permanently highlighted. This was usable, but
irritating. Your fix got Emacs back to how it was under WindowMaker. You have
saved a newbie!
In article <77d1pr$b2q$[EMAIL PROTECTED]>,
[EMAIL PROTECTED] (Brian Wheeler) wrote:
>
> Its not the virge driver, I don't think, because I haven't had virge troubles
> for years now.
>
> The problem may be if you're running a newish KDE. There's an option to
> "apply styles to non KDE applications"...this makes the fonts for emacs
> proportional, and makes it unusuable.
>
> sometimes, just running "xrdb ~/.Xdefaults" will fix it.
>
> Brian Wheeler
> [EMAIL PROTECTED]
>
> In article <[EMAIL PROTECTED]>,
> Paul Mackinlay <[EMAIL PROTECTED]> writes:
> > Has anyone experienced problems with an S3 ViRGE video card when using
> > KDE? On my computer it seems to have a problem with certain applications
> > (in particular emacs). All the font formatting is wrong and any
> > documents you load in emacs become illegible if you try and modify them.
> >
> > If anyone knows of S3 ViRGE driver problems or knows where to get
> > information about it please tell me. Thanks.
> >
> > Paul
> >
> > --
> > http://lark.ae.ic.ac.uk/~cmacki/PhD/
> >
> > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > C.P. Mackinlay, | E-Mail: [EMAIL PROTECTED]
> > Dept. of Aeronautics, | Tel: 0171 594 5110 (int. 45110)
> > Imperial College, |
> > Prince Consort Road, |
> > London SW7 2BZ |
> > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> >
> >
> >
>
============= Posted via Deja News, The Discussion Network ============
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
------------------------------
From: Franck ZOCCOLO <[EMAIL PROTECTED]>
Subject: numeric keypad and rxvt
Date: Sat, 23 Jan 1999 17:55:47 +0000
I'am sorry for this supid question, but I haven't found the answer
anywhere...
Using xterm, I'm able to use the numeric keypad of my keyboard : It
displays numbers when I press a key...
BUT
I'm unable to make it working correctly using rxvt (in fact, it's not
exactly rxvt, but wterm, a WindowMaker port of rxvt). For example, when
I press the "4" key, rxvt displays "q".
BUT
If I starts "man something", then the numeric keypad starts working !!!
I think I have to use an escape-sequence in order to make it working,
but I don't know which and how !!
Could someone help me ?
Please answer by e-mail : [EMAIL PROTECTED]
------------------------------
From: Todd Ostermeier <[EMAIL PROTECTED]>
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, 23 Jan 1999 13:04:46 -0600
On 23 Jan 1999, Larry Pyeatt wrote:
: In article <[EMAIL PROTECTED]>,
: Jay <[EMAIL PROTECTED]> writes:
: >
: > The only things I haven't found out (maybe someone can help):
: > - The maximum size of a C program and the maximum memory addressable by
: > said program (Like DOS was 640K unless DPMI was involved.)
:
: Ugh! Far pointers, several memory models, limits on how much you can
: malloc in one chunk... don't remind me.
:
: On Linux, you are only limited by the amount of RAM + swap space you have.
: Actually, I think there is a limit on the size of a single program,
: but that limit is measured in Gb.
Would that limit be 4GB (the maximum amount of memory addressable by a 32
pointer using byte-offsets)? Of course, if you *really* needed more than
4 GB of addressable memory, you could employ certain tricks (at a very low
level -- perhaps even down to the hardware :), such as using a word-offset
pointer, which would give you something like 16GB of addressable memory.
Don't ask me how to do this, however :)
________________________________
Todd Ostermeier
[EMAIL PROTECTED]
http://www.ews.uiuc.edu/~ostermer/index.html
ICQ UIN: 2253928
A-723
________________________________
------------------------------
** 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
******************************
