Linux-Development-Sys Digest #430, Volume #6 Thu, 25 Feb 99 22:14:20 EST
Contents:
Re: Edlin in Linux? (Was: Re: PROOF: Jesus *is* Lord of the Sabbath!) (Johan
Kullstam)
Re: Java 2 (Frank Sweetser)
Re: Problems compiling linux-2.2.2 (Michael Hirsch)
Any advantage to 2 swap partitions ? (Jim Cromie)
Re: Kernel 2.2 on top of 2.0.36 installations (bill davidsen)
How to kill process in state D? ([EMAIL PROTECTED])
Problem 'sync'ing Linux filesystem on i386 ([EMAIL PROTECTED])
Re: Here's A Truly UNIQUE Filesystem Problem !!! (bill davidsen)
Re: PROOF: Jesus *is* Lord of the Sabbath! (Jonathan Guyer)
Re: /dev/zero (Juergen Heinzl)
Re: Any advantage to 2 swap partitions ? (Peter Samuelson)
What happened to Simon Janes? (Ross Vandegrift)
Question about named pipes (FIFOs) ("David Sisk")
Re: looking for a way to stuff keystrokes (Eric Fagerburg)
Re: monolithic kernel and source that can only be compiled as a module (Peter
Samuelson)
Re: Edlin in Linux? (Was: Re: PROOF: Jesus *is* Lord of the Sabbath!) (Peter
Samuelson)
Re: Edlin in Linux? (Was: Re: PROOF: Jesus *is* Lord of the Sabbath!) (Johan
Kullstam)
----------------------------------------------------------------------------
Crossposted-To:
alt.society.underwear,comp.sys.mac.misc,comp.sys.amiga.hardware,fr.rec.voyages
Subject: Re: Edlin in Linux? (Was: Re: PROOF: Jesus *is* Lord of the Sabbath!)
From: Johan Kullstam <[EMAIL PROTECTED]>
Date: 25 Feb 1999 17:07:23 -0500
"Per Olsson" <[EMAIL PROTECTED]> writes:
> BTW, is there a Linux equivalent to DOS edlin? I'm asking just out of
> curiousity.
there's ed, but ed has more functionality if a somewhat more less
verbose user interface.
> Edlin is installed with Windows NT 4, why I dont know. Maybe to satisfy
> those used to edlin that find it hard to learn notepad.
notepad *is* hard. when ever i have to do much in notepad, it's about
all i can do to keep from hurling the computer out the window.
--
J o h a n K u l l s t a m
[[EMAIL PROTECTED]]
Don't Fear the Penguin!
------------------------------
From: Frank Sweetser <[EMAIL PROTECTED]>
Subject: Re: Java 2
Date: 25 Feb 1999 15:51:39 -0500
"Josh Toon" <[EMAIL PROTECTED]> writes:
> now that java 2 is open source, what does everyone think about more tightly
> integrating it into Linux for enterprise level app development?
why? what benefits does more tightly integrating it get?
--
Frank Sweetser rasmusin at wpi.edu fsweetser at blee.net | PGP key available
paramount.ind.wpi.edu RedHat 5.2 kernel 2.2.1 i586 | at public servers
It's possible that I'm just an idiot, and don't recognize a sleepy
slavemaster when I see one.
-- Larry Wall in <[EMAIL PROTECTED]>
------------------------------
From: Michael Hirsch <[EMAIL PROTECTED]>
Subject: Re: Problems compiling linux-2.2.2
Date: 25 Feb 1999 17:20:06 -0500
Usseglio Gaudi Francesco <[EMAIL PROTECTED]> writes:
> I had the same problem. I resolved it unchecking the "IP filtering" (or
> something similar:socket filter) and the problem disappeared. I think it's a
> bug in the 2.2.2 because the 2.2.1 works perfectly with that option on: i dig a
> little in the kernel source and i find that the function prototype of
> sk_run_filter has been changed. (btw i tried to patch it by i wasn't able to
> find where the sk_filter struct is defined).
> bye.
Thank, That fixed my compile problem, too. Definitely a kernel bug.
--
Michael D. Hirsch Work: (404) 727-7940
Emory University, Atlanta, GA 30322 FAX: (404) 727-5611
email: [EMAIL PROTECTED] http://www.mathcs.emory.edu/~hirsch/
Public key for encrypted mail available upon request (or finger
[EMAIL PROTECTED]).
------------------------------
From: Jim Cromie <[EMAIL PROTECTED]>
Subject: Any advantage to 2 swap partitions ?
Date: Thu, 25 Feb 1999 15:07:52 -0700
I just bought a new phat drive, and I want to spread my system
across both of them. Ive got RH 5.1 on a P2-233, a pair of
6 & 10-G UDMA IDE drives, both on same IDE interface,
CD-ROM on other.
Q: Is there any advantage to adding a swap partition to the 2nd drive ?
does Linux have the cleverness to use multiple swaps in an optimal way ?
Or is it just 'use the 1st till its full' then use the 2nd.
If Im not mistaken, EIDE cannot interleave requests to 2 drives, so the
value of all this optimization effort may be minimal, but what the
hell...
As an aside, is there any advantage/disadvantage to using 2 separate
1X-Meg swaps vs a single 2X-Meg one ?
If the latter is the case, would putting the new swap above the old one
in /etc/fstab cause the kernel to use it 1st ?
Is there anything in /proc that tells of swap(or other /dev/hdX)
utilization ?
(Im not at the box now).
Lastly, I justified this drive purchase as a fast, cheap, backup device.
Do any backup tools (midnite cmdr, etc..) have the smarts to remount a
backup partition as writable during the backup, and revert to readonly
as the default ?
------------------------------
From: [EMAIL PROTECTED] (bill davidsen)
Subject: Re: Kernel 2.2 on top of 2.0.36 installations
Date: 25 Feb 1999 23:07:24 GMT
In article <[EMAIL PROTECTED]>, mvrao <[EMAIL PROTECTED]> wrote:
| Can I just download the source for 2.2 and compile it ? Will it work
| with apps and sys utilities from 2.0.36 ?
Slackware 3.6 works fine, is missing only the ipchains stuff AFAIK. And
a quick look on a machine installed from the current snapshot (Jan 17)
shows that ipchains is there.
I believe Radhat has a bunch of rpms to upgrade 5.2, but I don't have
any solid info, as that's not what I hack.
--
bill davidsen <[EMAIL PROTECTED]> CTO, TMR Associates, Inc
Politicians and diapers have one thing in common. They should both be
changed regularly and for the same reason.
--Ted Symons(?)
------------------------------
From: [EMAIL PROTECTED]
Subject: How to kill process in state D?
Date: 25 Feb 1999 15:57:25 -0600
I had a tar process get an IO error writing to a IDE tape
drive. The process is now permanently in a wait state.
/proc show the status of the process to be in disk sleep.
Neither kill -15 nor kill -9 work. I've tried inserting and
removing tapes to try and cause an interrupt. No help.
Because this process has locked the tape drive, nothing else
can use it.
The process status looks like:
[/proc/21110]$ cat status
Name: tar
State: D (disk sleep)
Pid: 21110
PPid: 4831
Uid: 0 0 0 0
Gid: 0 0 0 0
VmSize: 900 kB
VmLck: 0 kB
VmRSS: 0 kB
VmData: 84 kB
VmStk: 16 kB
VmExe: 104 kB
VmLib: 648 kB
SigPnd: 00044100
SigBlk: 00000000
SigIgn: 00000000
SigCgt: 00000000
Is there a way to kill this process short of reboot?
Jim
--
// Jim Schulze [EMAIL PROTECTED] [EMAIL PROTECTED]
// finger [EMAIL PROTECTED] for pgp public key
BMC Software, Inc. makes no representations or promises
regarding the reliability, completeness, or accuracy of the
information provided in this discussion; all readers agree
not to rely on or take any action against BMC Software
in response to this information.
------------------------------
From: [EMAIL PROTECTED]
Subject: Problem 'sync'ing Linux filesystem on i386
Date: 25 Feb 1999 18:03:45 -0500
I have a P133, 6.3GB IBM IDE HD w/ 512KB Cache, and run RedHat 5.2 in last
1.5GB partition, 32MB ram.
The problem is that when I 'shutdown -h now' the machine, it halts on
'umount -a nsf' in /rc.0/K86???. I believe that is is the same problem that I
have when I try to 'sync' the drive at a prompt. The 'sync' process goes into
a D sleep/waiting mode, seen with 'ps aux' that can be run from another
terminal. I can even start X, run another sync, whatever, the number of dead
'sync' processes just add up. The hard drive light does flash when sync is run!
I even compiled a new kernel, the last thing of 'make zImage' is a 'sync', it
hung, I opened another window copied it to /boot and it worked just fine,
I thought it might be a Kernel problem, but no.
So... I guess that 'sync' is getting lost somewhere trying to exit?
My guesses and other observations...
How do I have Linux non-destructively verify the partition table
thoroughly? Dos,OS/2,NT,and Linux Fdisk don't report errors.
I can't seem to kill -9 sync... lowlevel I guess?
On bootup the Drive geometry is changed so that there are
764cyl,255hds,63spt = #sectors in BIOS, BUT!!!! The drive's cache is
reported to be 476Kb, NOT 512Kb???
fstab looks fine, I only have my linux drive mounted, I boot w/ System
Commander pointing to Lilo in my Linux Partition, but never had a booting
problem.
I can compile and do whatever, haven't had an error from not having a file
completely written to disk before I try to use it.
THIS HAS NOT ALWAYS BEEN THE CASE! That's whats most disturbing, shutdown
was fine for about 2 weeks, then this problem. When I startup, it then
reports that the Linux partition was not correctly shutdown and checks it.
I had a problem with my mounting my CDROM, it would hang too, but I
reinstalled and it hasn't cropped back up yet, is this related?
Using Kernel 2.0.3.??? what came w/ Redhat 5.2, should I get 2.2.???
Thanks for any assistance. This is my second try w/ Linux, the first
equally soured me, but I'd love to love it.
-John Labenski
[EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (bill davidsen)
Crossposted-To: comp.os.linux.help,comp.os.linux.questions,comp.os.linux
Subject: Re: Here's A Truly UNIQUE Filesystem Problem !!!
Date: 25 Feb 1999 23:16:06 GMT
In article <[EMAIL PROTECTED]>,
Dave Peticolas <[EMAIL PROTECTED]> wrote:
| On Thu, 25 Feb 1999 00:10:38 GMT, Carl Spalletta <[EMAIL PROTECTED]> wrote:
| >
| > Here's an interesting problem: How can a file with ZERO links remain
| >listed in an ext2 filesystem???
| >
| >$ ls -li /lib/libc.so.5
| >EXT2-fs warding (device 03:42): ext2_free_inode: bit already cleared
| >for inode 29477
| > 29477 ?--------- 0 root root 0 Jan 1 1970 libc.so.5
|
| I don't know enough about ext2 to know exactly what's going on here,
| but have you tried using /sbin/debugfs to remove the file? I have used
| that tool to remove bizarre files created by a corrupt filesystem.
The directory entry provides a pointer to an inode, the inode provides
the info. At the rick of sounding really dumb, have you tried just using
'rm' of this. You have a directory entry botch.
I don't see why that would be a problem. Debugfs is powerful, I've used
it when the inode was botched, but not for the directory.
--
bill davidsen <[EMAIL PROTECTED]> CTO, TMR Associates, Inc
Politicians and diapers have one thing in common. They should both be
changed regularly and for the same reason.
--Ted Symons(?)
------------------------------
Date: Thu, 25 Feb 1999 19:08:52 -0500
From: [EMAIL PROTECTED] (Jonathan Guyer)
Crossposted-To:
alt.society.underwear,comp.sys.mac.misc,comp.sys.amiga.hardware,fr.rec.voyages
Subject: Re: PROOF: Jesus *is* Lord of the Sabbath!
In article <[EMAIL PROTECTED]>,
[EMAIL PROTECTED] wrote:
> On Wed, 24 Feb 1999 08:10:43 -0800, Pablo <[EMAIL PROTECTED]> wrote:
>
> >I wonder what kind of Mac God uses?
>
> The Amiga is GOD!
God is dead? 8^)
--
Jonathan E. Guyer
<http://www.his.com/~jguyer/>
------------------------------
From: [EMAIL PROTECTED] (Juergen Heinzl)
Subject: Re: /dev/zero
Date: Fri, 26 Feb 1999 00:57:12 GMT
In article <7b4lf0$mq8$[EMAIL PROTECTED]>, John McKown wrote:
>It's real neat for initializing a new file to 0x00!
>
>dd if=/dev/zero of=myfile bs=512 count=20
>Will create a 10K file which has 0x00 as every byte.
I always dreamt of that 8) ... nah, but serious again, /dev/zero
is often used for memory mapped files for instance. Just remove it ... NO ...
just kidding since without /dev/zero no loading of shared libraries anymore,
at least with the libc5 loader (the libc6 one uses /dev/null (?)).
Cheers,
Juergen
--
\ Real name : J�rgen Heinzl \ no flames /
\ EMail Private : [EMAIL PROTECTED] \ send money instead /
\ Phone Private : +44 181-332 0750 \ /
------------------------------
From: [EMAIL PROTECTED] (Peter Samuelson)
Subject: Re: Any advantage to 2 swap partitions ?
Date: 25 Feb 1999 18:51:30 -0600
Reply-To: Peter Samuelson <[EMAIL PROTECTED]>
[Jim Cromie <[EMAIL PROTECTED]>]
> I just bought a new phat drive, and I want to spread my system across
> both of them. Ive got RH 5.1 on a P2-233, a pair of 6 & 10-G UDMA
> IDE drives, both on same IDE interface, CD-ROM on other.
> If Im not mistaken, EIDE cannot interleave requests to 2 drives, so
> the value of all this optimization effort may be minimal, but what
> the hell...
You are not mistaken. What you really want to do (I do mean *really*)
is put the two drives on *separate* channels, rather than as master and
slave of a single channel. Then put the CDROM with whichever hard
drive will get less traffic.
A coworker tells the story of a factory-installed NT box he had to
troubleshoot once -- it had two IDE drives and was set up to have them
mirror each other. The HD access was dog slow. Clearly some moron at
Tangent Corp. had, you guessed it, put both drives on one channel. My
friend figured this out but management wouldn't let him open up the
case -- instead they called Tangent tech support, who were (as usual)
without clue, even when they sent someone here to have a look! Finally
my friend got permission to do his thing and of course the box behaved
normally as soon as he switched the channels around.
This would only surprise anyone who has never dealt with Tangent, BTW.
> Q: Is there any advantage to adding a swap partition to the 2nd
> drive? does Linux have the cleverness to use multiple swaps in an
> optimal way?
I think so, as long as you set the same swap priority. (Use a "pri=X"
option in /etc/fstab.) To quote swapon(2): "Higher numbers mean higher
priorities." Linux will fill up an entire higher-priority swap space
before touching a lower-priority space.
> Is there anything in /proc that tells of swap(or other /dev/hdX)
> utilization ? (Im not at the box now).
/proc/swaps
/proc/meminfo
/proc/partitions
> Do any backup tools (midnite cmdr, etc..) have the smarts to remount
> a backup partition as writable during the backup, and revert to
> readonly as the default ?
Dunno. Not hard to write a three-line wrapper script, if not....
--
Peter Samuelson
<sampo.creighton.edu!psamuels>
------------------------------
From: Ross Vandegrift <[EMAIL PROTECTED]>
Subject: What happened to Simon Janes?
Date: Thu, 25 Feb 1999 17:53:06 -0500
Reply-To: [EMAIL PROTECTED]
Anyone know where Simon Janes went? I sent him an email about some
stuff in the eql driver, and his listed email bounced. I would search
DejaNews, but that's.... that's like stalking.
--
Ross Vandegrift | Eric J. Fenderson
ATTENTION: I have **finally** gotten my permit!!!
------------------------------
From: "David Sisk" <[EMAIL PROTECTED]>
Subject: Question about named pipes (FIFOs)
Date: Fri, 26 Feb 1999 20:24:13 -0500
This is a "synchronization" type question.
Say you have Process1 which is listening on a named pipe (or an unnamed
pipe, for that matter) called Pipe1. Process2 writes '123' to Pipe1.
Process3 writes 'ABC' to Pipe1. Process2 and Process3 attempt to do this at
the exact same time. What happens?
1) Process1 gets either '123ABC' or 'ABC123'. In other words, only one
process at a time can write into the pipe, and it will be the first process
that gets there.
OR
2) Process1 might get '1A23BC', 'AB123C', etc. In other words, Process1
receives the data in whatever order it happens to get placed in the pipe by
whatever writing process.
Which scenario is correct?
Also, thanks to all you folks for answering many of my questions. I'm
trying to get comfortable with unix and unix programming, and you folks are
really helpful. It is much appreciated!
Please email as well as posting, if you can and don't mind.
Regards,
--
David C. Sisk
The Unofficial ORACLE on NT site
http://www.ipass.net/~davesisk/oont.htm
------------------------------
From: Eric Fagerburg <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.development.apps
Subject: Re: looking for a way to stuff keystrokes
Date: Thu, 25 Feb 1999 19:09:52 -0700
Willem Rein Oudshoorn wrote:
> take a look at the
>
> sti.c
>
> program in the contrib directory ...
Many thanks! TIOCSTI is just what I was looking for. About the only
thing I can't figure out how to get to work is Shift-PgUp and
Shift-PgDn.
> I don't know why you should want this,
This is for an application that allows a help desk person to view a
user's screen and even type commands while the user watches so they see
how a problem is resolved.
Thanks again,
Eric Fagerburg
[EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Peter Samuelson)
Subject: Re: monolithic kernel and source that can only be compiled as a module
Date: 25 Feb 1999 19:23:22 -0600
Reply-To: Peter Samuelson <[EMAIL PROTECTED]>
[Tomi Ollila <[EMAIL PROTECTED]>]
> I am compiling a kernel, and in a many reasons I want to be
> monolithic, i.e. no support for modules, but all needed stuff
> compiled as "resident".
Out of curiosity, why? I don't doubt you have a good reason, mind.
Anyway...
> Now I noticed that some of the source code can be compiled as a
> "module" only.
> What is the reason(s) for that.
Probably because it's easier to write a module than builtin code, so
some authors never did do the integration part. It's easier because
you don't have to figure out where to call the init function from other
kernel code, and also easier because you're probably already debugging
the code as a module (`rmmod foo;insmod foo' is a *lot* more convenient
than `reboot').
> Would there be a nice way to "convert" some of the code so that those
> could be compiled in -- or a nasty ad-hoc hack way if that is not
> possible.
Look at existing source for examples. There are plenty. In the
simplest case you just have to call init_module() from some kernel
function that's already compiled in -- there should be a file for that
purpose close by (in drivers/net/ it's Space.c, for example). Then
edit the Makefile in that directory, and the nearest Config.in file.
Also if the module takes parameters you may have to do a little magic
to get it to use the kernel command line.
--
Peter Samuelson
<sampo.creighton.edu!psamuels>
------------------------------
From: [EMAIL PROTECTED] (Peter Samuelson)
Crossposted-To:
alt.society.underwear,comp.sys.mac.misc,comp.sys.amiga.hardware,fr.rec.voyages
Subject: Re: Edlin in Linux? (Was: Re: PROOF: Jesus *is* Lord of the Sabbath!)
Date: 25 Feb 1999 19:33:51 -0600
Reply-To: Peter Samuelson <[EMAIL PROTECTED]>
[Per Olsson <[EMAIL PROTECTED]>]
> > BTW, is there a Linux equivalent to DOS edlin? I'm asking just out
> > of curiousity.
[Johan Kullstam <[EMAIL PROTECTED]>]
> there's ed, but ed has more functionality if a somewhat more less
> verbose user interface.
The question really should have been "is there a DOS equivalent to Unix
ed?" AFAICT edlin was really an ed clone but of course with most of
the useful functionality taken out.
> > Edlin is installed with Windows NT 4, why I dont know. Maybe to
> > satisfy those used to edlin that find it hard to learn notepad.
edlin *is* scriptable, I believe (insofar as anything from DOS or
Windoze could be called scriptable). That would be a reason not to
take it out....
...though IIRC they *did* rip it out of DOS 5 or 6, in favor of
EDIT.COM. Maybe nobody will believe me but I was actually upset, since
I ran a machine slow enough that EDIT.COM took awhile to come up (it's
actually just a wrapper for the QBASIC.EXE builtin editor). Same
reason I sometimes use vi when I have XEmacs available here....
--
Peter Samuelson
<sampo.creighton.edu!psamuels>
------------------------------
Crossposted-To:
alt.society.underwear,comp.sys.mac.misc,comp.sys.amiga.hardware,fr.rec.voyages
Subject: Re: Edlin in Linux? (Was: Re: PROOF: Jesus *is* Lord of the Sabbath!)
From: Johan Kullstam <[EMAIL PROTECTED]>
Date: 25 Feb 1999 21:06:17 -0500
[EMAIL PROTECTED] (Peter Samuelson) writes:
> ...though IIRC they *did* rip it out of DOS 5 or 6, in favor of
> EDIT.COM. Maybe nobody will believe me but I was actually upset, since
> I ran a machine slow enough that EDIT.COM took awhile to come up (it's
> actually just a wrapper for the QBASIC.EXE builtin editor). Same
> reason I sometimes use vi when I have XEmacs available here....
heh. when things get too horrible for emacs, i drop into line-by-line
mode and use ed. ed really isn't all that bad if you know how to work
it. its small size is good for rescue disks. and, perhaps most
importantly, ed will not corrupt your precious bodily fluids! ;-)
--
J o h a n K u l l s t a m
[[EMAIL PROTECTED]]
Don't Fear the Penguin!
------------------------------
** 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
******************************
