Linux-Development-Sys Digest #978, Volume #6 Tue, 20 Jul 99 23:13:46 EDT
Contents:
Re: when will Linux support > 2GB file size??? (bill davidsen)
Re: dual Celeron MB blows up constantly! (bill davidsen)
Linux on PS/2 MCA ESDI???? (User Bsdbob BSD Bob)
Re: How to read/write/format MS-DOS floppy disks??? (Peter Pointner)
Re: Server linux config ([EMAIL PROTECTED])
Re: Serial device (c language) (bill davidsen)
Re: Server linux config (Warren Young)
Re: NCR 53C710 Fast SCSI-2 Controller (bill davidsen)
Re: How to use sockets? (Warren Young)
Re: Linux on PS/2 MCA ESDI???? (Roy Grimm)
Re: Riva TNT Drivers (Andrew Lipnitsky)
libc/glibc and gcc/egcs issues (Lou Grinzo)
Re: Memory fault or broken kernel?? (Andi Kleen)
Re: Linux on PS/2 MCA ESDI???? (Robert Lynch)
Re: MICROSOFT LINUX DISTRIBUTION (Christopher Browne)
JavaWebServer1.1.3 in linux (Matt)
Re: Kernel--User (Frank v Waveren)
Kernel--User (cwu)
Re: S3 Savage4 and RedHat6 (Germ)
Re: libc/glibc and gcc/egcs issues (Paul Kimoto)
Re: libc/glibc and gcc/egcs issues (David T. Blake)
DOS FS module problem in 2.3.10 kernel? (Andrea Dixon)
Re: when will Linux support > 2GB file size??? (Christopher B. Browne)
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (bill davidsen)
Crossposted-To: comp.os.linux.advocacy
Subject: Re: when will Linux support > 2GB file size???
Date: 20 Jul 1999 18:38:03 GMT
In article <7mtejn$2nee$[EMAIL PROTECTED]>,
Leslie Mikesell <[EMAIL PROTECTED]> wrote:
| In article <7mo5sf$3678$[EMAIL PROTECTED]>,
| bill davidsen <[EMAIL PROTECTED]> wrote:
| >I don't see any easy fix for mmap(), any fix would include 64 bit
| >subscripts in many places, and I think that *would* be a performance hit
| >of vast impact.
|
| Until (not unless...) you have a box with more than 2 gigs of RAM.
| Then the performance hit will be that you can't use all the RAM.
Yes, but this is only a problem with more than 2GB per process, the
memory management can certainly allow more than 2GB physical to be used
even if any given process will only get 2GB.
I see mmap() as a real problem, because there's no reasonable way to
have a subscript become type offset_t instead of int and not have an
impact on the source code. I can envision that by using a magic compiler
you could cover the conversion, but people doing computations of
subscripts might find problems.
You could simply make int 64 bits in the compiler and let the overhead
fall where it might, that offers a path to not having source code
changes, I think. This is an ugly problem, and the cleanest solution may
be to leave mmap() alone and just change seek and tell calls.
--
bill davidsen <[EMAIL PROTECTED]> CTO, TMR Associates, Inc
The Internet is not the fountain of youth, but some days it feels like
the fountain of immaturity.
------------------------------
From: [EMAIL PROTECTED] (bill davidsen)
Subject: Re: dual Celeron MB blows up constantly!
Date: 20 Jul 1999 19:27:50 GMT
In article <[EMAIL PROTECTED]>,
Brian Gilman <[EMAIL PROTECTED]> wrote:
| Well, I go my abit dual celeron board today and have had nothing but
| problems with different kernels.....I wanted to use this board to learn
| about smp and programming threads with smp but, it's just not stable
| enough.....Sigh......Does anyone know what kernel version is considered
| the most *stable* for smp? Thanks in advance!
Is it stable running a uniprocessor kernel?
Have you swapped CPUs and repeated running uniprocessor?
Are you adding overclocking to SMP operation in the list of unsupported
(by Intel) conditions?
I would start running this board at rated speed uning one CPU, and if
that's stable swap CPUs and repeat. Then if you're doing o/c, up the
speed and repeat. Only if all that is stable should you try SMP.
I've been running SMP systems since July 4 1996, and haven't seen any
2.2 kernel which was unstable to the extent you imply. Or even unstable
to the point where I have had a problem in production.
--
bill davidsen <[EMAIL PROTECTED]> CTO, TMR Associates, Inc
The Internet is not the fountain of youth, but some days it feels like
the fountain of immaturity.
------------------------------
From: User Bsdbob BSD Bob <[EMAIL PROTECTED]>
Subject: Linux on PS/2 MCA ESDI????
Date: 20 Jul 1999 19:51:14 GMT
Anyone have a pointer to a Linux that is known to boot and run
directly on an IBM PS/2 Model 80, 10mb ram, 2 x 315mb ESDI drives?
If so, I would like the pointer to it.
Thanks
Bob Keys
------------------------------
From: Peter Pointner <[EMAIL PROTECTED]>
Subject: Re: How to read/write/format MS-DOS floppy disks???
Date: Tue, 20 Jul 1999 19:09:28 GMT
GUAY Dominic <[EMAIL PROTECTED]> wrote:
> Hi.
> I need to build a system with support for MS-DOS floppies under Linux. Do
> any of you know of any good software to do this?
> thanks,
> D
mtools. no need to mount, simply
mformat a:
mcopy a:...
...
Low level formatting can be done with fdformat ("util"-package at least
on my machine).
And the loopback device can be nice:
dd if=/dev/fd0 of=floppyimage
mount -o loop -t vfat floppyimage /mnt
rm /mnt/oldstuff
cp newstuff /mnt
umount /mnt
dd if=floppyimage of=/dev/fd0
(Warning: be carefull when using dd, especially as root. It's powerfull
and unforgiving)
Peter
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: Server linux config
Date: Mon, 19 Jul 1999 23:47:45 -0700
Reply-To: [EMAIL PROTECTED]
bill davidsen <[EMAIL PROTECTED]> wrote:
> I'm trying to use 2.2.6 as a server, and it doesn't seem to be able to
> support more than a few hundred processes. I have made the obvious
> changes in linux/task.h (# proc) and linux/limits.h (# of open files),
> and it still appears to be running out of file descriptors.
are there ulimits set?
-ckm
------------------------------
From: [EMAIL PROTECTED] (bill davidsen)
Subject: Re: Serial device (c language)
Date: 20 Jul 1999 20:40:52 GMT
In article <[EMAIL PROTECTED]>,
root <[EMAIL PROTECTED]> wrote:
| - MIN = 0 and TIME > 0 (TIME = 500) : it doesn't work, ie I use the
| read function, there is nothing to read, the function return immediatly
| (there is no timeout )
That should time out, although the timeout will be fairly long if I
remember correctly. The TIME is in sec, or 1/10ths, or something large,
not ms. Are you sure you waited long enough?
I haven't used this recently, so this is from memory.
--
bill davidsen <[EMAIL PROTECTED]> CTO, TMR Associates, Inc
The Internet is not the fountain of youth, but some days it feels like
the fountain of immaturity.
------------------------------
From: Warren Young <[EMAIL PROTECTED]>
Subject: Re: Server linux config
Date: Tue, 20 Jul 1999 15:37:46 -0600
bill davidsen wrote:
>
> I'm trying to use 2.2.6 as a server, and it doesn't seem to be able to
> support more than a few hundred processes. I have made the obvious
> changes in linux/task.h (# proc) and linux/limits.h (# of open files),
> and it still appears to be running out of file descriptors.
>
> I didn't have this problem with 2.0 kernels, they ran when patched. What
> has changed?
Have you changed the ulimits for the user that owns the server
processes? By default, no user can create as many processes as the
system can support. Ditto for file handles. It's to guard against
denial-of-service attacks.
--
= Warren -- See the *ix pages at http://www.cyberport.com/~tangent/ix/
=
= ICBM Address: 36.8274040 N, 108.0204086 W, alt. 1714m
------------------------------
From: [EMAIL PROTECTED] (bill davidsen)
Crossposted-To: comp.os.linux.hardware
Subject: Re: NCR 53C710 Fast SCSI-2 Controller
Date: 20 Jul 1999 19:08:21 GMT
In article <[EMAIL PROTECTED]>,
Johan Kullstam <[EMAIL PROTECTED]> wrote:
| [EMAIL PROTECTED] (bill davidsen) writes:
| > I would try ncr53c8xx before going into panic mode. Also note that most
| > computer shows will provide you with a supported controller for <$50,
| > either a generic NCR (if PCI) or a AHA154x if ISA.
|
| they will? all the computer shows i've been to have never heard
| anything non-adaptec when it comes to scsi adapter cards.
Well, around Albany NY, DC, and NYC they will. Online places certainly
have adaptors for cheap. YMMV.
--
bill davidsen <[EMAIL PROTECTED]> CTO, TMR Associates, Inc
The Internet is not the fountain of youth, but some days it feels like
the fountain of immaturity.
------------------------------
From: Warren Young <[EMAIL PROTECTED]>
Subject: Re: How to use sockets?
Date: Tue, 20 Jul 1999 15:35:50 -0600
[EMAIL PROTECTED] wrote:
>
> How do I populate the "struct sockaddr" used in the bind() call? I have
> looked around without much success. I would like to connect the socket
> to a IP port (e.g. 123.45.6.78:9090).
If this is a client program (i.e. one making the connection), you don't
need to call bind(). Just call connect(), and it will attach to an
appropriate interface and local port.
If this is a server program (i.e. one waiting for connections), you
usually don't need to bind to a particular address: use zeroes for the
address, and put only your port into the structure.
Note that struct sockaddr is not the structure you use most of the
time. Instead, you use sockaddr_in, whose contents should be much more
understandable. Both structures are the same size, just with different
layouts.
See my FAQ (linked below) for pointers. It's for Winsock, but it's
close enough to BSD sockets that the information will be helpful to
you. Also, in the FAQ's Resources section, under Web Pages, you'll find
pointers to other FAQs, including at least one BSD sockets FAQ.
Good luck,
--
= Warren Young, maintainer of the Winsock Programmer's FAQ at:
= http://www.cyberport.com/~tangent/programming/winsock/
=
= ICBM Address: 36.8274040 N, 108.0204086 W, alt. 1714m
------------------------------
From: Roy Grimm <[EMAIL PROTECTED]>
Subject: Re: Linux on PS/2 MCA ESDI????
Date: Tue, 20 Jul 1999 16:27:32 -0500
User Bsdbob BSD Bob wrote:
>
> Anyone have a pointer to a Linux that is known to boot and run
> directly on an IBM PS/2 Model 80, 10mb ram, 2 x 315mb ESDI drives?
> If so, I would like the pointer to it.
>
> Thanks
>
> Bob Keys
If I'm not mistaken, Linux does not support microchannel at all. Go to
www.linux.org and look for the supported architectures...
Roy
------------------------------
From: Andrew Lipnitsky <[EMAIL PROTECTED]>
Subject: Re: Riva TNT Drivers
Date: Wed, 21 Jul 1999 00:54:11 +0300
Hello Ian.
Ian Murphy wrote:
> Does anybody know of or possess updated drivers for the Riva TNT chipset for
> Linux, as I'm having trouble getting X-Windows to recognise my Creative Labs
> Graphics Blaster TNT.
Please, take a look to
ftp://ftp2.detonator.nvidia.com/pub/drivers/linux
and http://www.nvidia.com
You will find what you are lookinng for.
Andrew.
------------------------------
From: [EMAIL PROTECTED] (Lou Grinzo)
Subject: libc/glibc and gcc/egcs issues
Date: Tue, 20 Jul 1999 18:01:11 -0400
I've found a number of sites on the Web that address
installing various versions of the C lib and
switching between them, but I've yet to find any
concrete description of exactly why we're seeing
incompatibilities in this area.
I know that the StarOffice situation was caused by
the SD developers doing thing they shouldn't under
glibc 2.0, which caused SO to break under 2.1, but
that's not what I'm talking about.
There's a white paper on RedHat's site that says
that if you build a "hello, world" program under
RH 6.0 without statically linking the C library,
and then run it under RH 5.2, it will seg fault.
Is this because of a change to the library itself,
or the tools used to build it (such as gcc vs.
egcs)? What exactly changed? Is it a difference
in parameter passing, a change in the library
format, or what?
Also, why was this change made? I assume there was
a good reason, but I'm curious as to the details,
and I can't find an explanation.
I would appreciate some help, particularly in the
form of URL's with "official" explanations.
Lou
------------------------------
From: Andi Kleen <[EMAIL PROTECTED]>
Subject: Re: Memory fault or broken kernel??
Date: 20 Jul 1999 23:40:14 +0200
Michael Huhn <[EMAIL PROTECTED]> writes:
> Hello out there,
> a collegue of mine is running a linx-box with kernel 2.0.35.
> After a strange system-crash due to a strange network-problem the system
> behaves very strange.
[...]
Sounds like a hardware problem.
-Andi
--
This is like TV. I don't like TV.
------------------------------
Date: Tue, 20 Jul 1999 15:32:46 -0700
From: Robert Lynch <[EMAIL PROTECTED]>
Subject: Re: Linux on PS/2 MCA ESDI????
Roy Grimm wrote:
>
> User Bsdbob BSD Bob wrote:
> >
> > Anyone have a pointer to a Linux that is known to boot and run
> > directly on an IBM PS/2 Model 80, 10mb ram, 2 x 315mb ESDI drives?
> > If so, I would like the pointer to it.
> >
> > Thanks
> >
> > Bob Keys
>
> If I'm not mistaken, Linux does not support microchannel at all. Go to
> www.linux.org and look for the supported architectures...
>
> Roy
This is not correct. At one time there was a web-site with all kinds of
MCA stuff. However, with the 2.2 kernel, I believe this is all just
rolled into Linux.
Read linux/Documentation/mca.txt for details.
Bob L.
--
Robert Lynch-Berkeley CA [EMAIL PROTECTED]
http://www.best.com/~rmlynch/
------------------------------
From: [EMAIL PROTECTED] (Christopher Browne)
Subject: Re: MICROSOFT LINUX DISTRIBUTION
Reply-To: [EMAIL PROTECTED]
Date: Tue, 20 Jul 1999 23:29:30 GMT
On 20 Jul 1999 07:19:31 +0100, Nix <$}xinix{[email protected]> wrote:
>"ELSID Software Systems LTD." <[EMAIL PROTECTED]> writes:
>
>> All they have to do is bring out a is bring out a Linux like os that
>> interfaces
>> with all the current Linux loadable module specifications as well at
>> their own, call it
>
>There aren't any `loadable module specifications'.
>
>The only `specifications' are
>
>- the source, and
>
>- a promise that nothing whatsoever will be done to avoid breaking
> binary compatibility between kernel versions.
>
>Linus (rightly) doesn't like pandering to the restrictive needs of
>binary-only modules overmuch.
I think you need to amend that second specification to:
"a promise that binary compatibility won't *often* be broken between
kernel versions."
I seem to recall him indicating that it was considered legitimate to
occasionally break compatibility. Consider that:
- If the code is available in source form, this doesn't break.
- If the code isn't available in source form, authors of binary-only
modules should periodically reexamine whether the module continues
to be relevant, or whether it continues to need to remain
proprietary.
If someone wants to persistently release proprietary modules, there is
no assigned mandate to make this stable in the long term.
--
"...Unix, MS-DOS, and Windows NT (also known as the Good, the Bad, and
the Ugly)." -- Matt Welsh
[EMAIL PROTECTED] <http://www.ntlug.org/~cbbrowne/linuxkernel.html>
------------------------------
From: Matt <[EMAIL PROTECTED]>
Crossposted-To: comp.lang.java.help,comp.os.linux.misc,comp.os.linux.development.apps
Subject: JavaWebServer1.1.3 in linux
Date: Wed, 21 Jul 1999 00:26:46 +0100
Hi,
Has anyone managed to start the sun JavaWebServer1.1.3 in Linux
using the jdk1.2 yet if so how is it done ?
Many thanks
Matt
------------------------------
From: [EMAIL PROTECTED] (Frank v Waveren)
Subject: Re: Kernel--User
Crossposted-To: comp.os.linux.networking,redhat.kernel.general
Date: Wed, 21 Jul 1999 01:43:25 GMT
In article <[EMAIL PROTECTED]>,
cwu <[EMAIL PROTECTED]> writes:
> Hi, all:
>
> How can I get info from Kernel; and how can I set parameters to kernel. I
> am on the project to modify Linux kernel (add some functions in linux
> kernel). I want the functions can be trigger from user domain ( I can set
> the parameters on the user domain and communictate with kernel, finally
> these parameters will reach the functions in linux kernel). and then I want
> to get the result(report) from kernel to user domain.
>
> Is it possible? How?
>
> Anyone can give me the hint or any pointer to study this kind of
> user--kernel communication?
>
> Thanks,
>
> cwu
>
>
>
> ------------------ Posted via SearchLinux ------------------
> http://www.searchlinux.com
I'd think the /proc filesystem would be ideal for what you are looking for.
see /usr/src/linux/Documentation/proc.txt
BTW, isn't it ironic that posts via searchlinux always come in CRLF format?
You'd think they'd change that If they wanted to be taken seriously.
--
Frank v Waveren
[EMAIL PROTECTED]
ICQ# 10074100
------------------------------
From: cwu <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.networking,redhat.kernel.general
Subject: Kernel--User
Date: Wed, 21 Jul 1999 00:31:09 GMT
Hi, all:
How can I get info from Kernel; and how can I set parameters to kernel. I
am on the project to modify Linux kernel (add some functions in linux
kernel). I want the functions can be trigger from user domain ( I can set
the parameters on the user domain and communictate with kernel, finally
these parameters will reach the functions in linux kernel). and then I want
to get the result(report) from kernel to user domain.
Is it possible? How?
Anyone can give me the hint or any pointer to study this kind of
user--kernel communication?
Thanks,
cwu
================== Posted via SearchLinux ==================
http://www.searchlinux.com
------------------------------
From: Germ <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.development.apps,comp.os.linux.x,comp.os.linux.hardware
Subject: Re: S3 Savage4 and RedHat6
Date: Wed, 21 Jul 1999 01:27:40 GMT
> Carlos Rodrigues <[EMAIL PROTECTED]> wrote:
> Hi. I'm thinking of buying a Savage4 based graphics card and I would >
like to
> know if someone tried it with RedHat6. Any comments would be
> appreciated.
I got it working with an unsupported driver from Creative, but have yet
to get it out of 800x600...
You can get the driver here:
http://support.soundblaster.com/files/download.asp?OS=Beta&prod=db_savag
e4
My honest opinion though is to get a different card, I would have if I
had checked. </Germ>
--
function life()
if $you can't `grep` * {
| /dev/null;
}
Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
------------------------------
From: [EMAIL PROTECTED] (Paul Kimoto)
Subject: Re: libc/glibc and gcc/egcs issues
Date: 20 Jul 1999 21:42:46 -0500
Reply-To: [EMAIL PROTECTED]
In article <MPG.11febbb6b063d3219896f5@news-server>, Lou Grinzo wrote:
> I know that the StarOffice situation was caused by
> the SD developers doing thing they shouldn't under
> glibc 2.0, which caused SO to break under 2.1, but
> that's not what I'm talking about.
>
> There's a white paper on RedHat's site that says
> that if you build a "hello, world" program under
> RH 6.0 without statically linking the C library,
> and then run it under RH 5.2, it will seg fault.
>
> Is this because of a change to the library itself,
> or the tools used to build it (such as gcc vs.
> egcs)? What exactly changed?
Someone else will have to address what _exactly_ changed,
but RH 6.0 is a glibc-2.1 system, and RH 5.X is a glibc-2.0
system. According to the usual convention, executables
linked against glibc-2.1 can be run using glibc-2.x, where
x >= 1.
--
Paul Kimoto <[EMAIL PROTECTED]>
------------------------------
From: [EMAIL PROTECTED] (David T. Blake)
Subject: Re: libc/glibc and gcc/egcs issues
Date: 21 Jul 1999 01:33:19 GMT
Reply-To: [EMAIL PROTECTED]
Lou Grinzo <[EMAIL PROTECTED]> wrote:
> I've found a number of sites on the Web that address
> installing various versions of the C lib and
> switching between them, but I've yet to find any
> concrete description of exactly why we're seeing
> incompatibilities in this area.
>
> I know that the StarOffice situation was caused by
> the SD developers doing thing they shouldn't under
> glibc 2.0, which caused SO to break under 2.1, but
> that's not what I'm talking about.
>
> There's a white paper on RedHat's site that says
> that if you build a "hello, world" program under
> RH 6.0 without statically linking the C library,
> and then run it under RH 5.2, it will seg fault.
The glibc 2.1 documentation has answers to all of your
questions.
--
Dave Blake
[EMAIL PROTECTED]
------------------------------
From: Andrea Dixon <[EMAIL PROTECTED]>
Subject: DOS FS module problem in 2.3.10 kernel?
Date: Wed, 21 Jul 1999 02:07:22 GMT
I just downloaded the 2.3.10 kernel.
Built, installed, and reboot.
When modprobe tries to load the FAT module for access
to my (crappy) FAT partition, it complains that
symbol "update_vm_cache" is not found.
Also, the bt848 video module cannot find "init_mm".
Source for both modules is included in the 2.3.10 download!
Any thoughts on this matter? Thanks!
Jamie Strachan
------------------------------
From: [EMAIL PROTECTED] (Christopher B. Browne)
Crossposted-To: comp.os.linux.advocacy
Subject: Re: when will Linux support > 2GB file size???
Reply-To: [EMAIL PROTECTED]
Date: Wed, 21 Jul 1999 02:31:04 GMT
On 20 Jul 1999 18:38:03 GMT, bill davidsen <[EMAIL PROTECTED]> posted:
>In article <7mtejn$2nee$[EMAIL PROTECTED]>,
>Leslie Mikesell <[EMAIL PROTECTED]> wrote:
>| In article <7mo5sf$3678$[EMAIL PROTECTED]>,
>| bill davidsen <[EMAIL PROTECTED]> wrote:
>
>| >I don't see any easy fix for mmap(), any fix would include 64 bit
>| >subscripts in many places, and I think that *would* be a performance hit
>| >of vast impact.
>|
>| Until (not unless...) you have a box with more than 2 gigs of RAM.
>| Then the performance hit will be that you can't use all the RAM.
>
>Yes, but this is only a problem with more than 2GB per process, the
>memory management can certainly allow more than 2GB physical to be used
>even if any given process will only get 2GB.
>
>I see mmap() as a real problem, because there's no reasonable way to
>have a subscript become type offset_t instead of int and not have an
>impact on the source code. I can envision that by using a magic compiler
>you could cover the conversion, but people doing computations of
>subscripts might find problems.
>
>You could simply make int 64 bits in the compiler and let the overhead
>fall where it might, that offers a path to not having source code
>changes, I think. This is an ugly problem, and the cleanest solution may
>be to leave mmap() alone and just change seek and tell calls.
I suggest that this is a *common* issue that hits:
a) Memory mapping,
b) File mapping, and
c) Date mapping.
All three have roughly the same sorts of limitations that result from
32 bit values, and you hit all three in the FILE * structure.
There are effectively three choices:
a) Leave it alone, letting 32 bit systems have 32 bit limitations, and
64 bit systems have 64 bit limits, which is the situation we have at
present,
b) Move to 64 bit representations on all platforms, which results in a
performance "hit" for 32 bit systems, but which has the positive effect
that it is a "linear" change, or
c) Segment things, having the *option* of using a 64 bit API when
desired, and a 32/64 bit API (platform-dependent) when you "don't care."
My suspicion is that the best way to handle this is to go with option a)
for the time being, and then move to b) in about five years when 64 bit
processors become dominant in the marketplace.
--
Rules of the Evil Overlord #5: "The artifact which is the source of my
power will not be kept on the Mountain of Despair beyond the River of
Fire guarded by the Dragons of Eternity. It will be in my safe-deposit
box. The same applies to the object which is my one weakness."
[EMAIL PROTECTED] <http://www.ntlug.org/~cbbrowne/linuxkernel.html>
------------------------------
** 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
******************************
