Linux-Development-Sys Digest #10, Volume #7 Fri, 30 Jul 99 11:13:54 EDT
Contents:
Re: Linux Journal - worth or not? (Harald Arnesen)
Call for Authors: SOHO System Admin Guide (Gary Lawrence Murphy)
Re: Writing shared libraries (Karl Heyes)
Re: Linux SCSI Performance Issues (Douglas Gilbert)
The Lex & Yacc source code for C++ (jievis)
Re: How to find Network interface index? (Andreas Jaeger)
Re: The Lex & Yacc source code for C++ (Adam Przybyla)
JFS for linux? (Stefan Opperskalski)
Re: Linux Journal - worth or not? ("Chris Liddell")
Re: Compiling Linux 1.0 (Matthias Maisenbacher)
Re: The Lex & Yacc source code for C++ ("Noah Roberts (jik-)")
Re: glibc 2.1.1 clock() problem (returning -ve number?) (Dr H. T. Leung)
Re: JFS for linux? (J.H.M. Dassen (Ray))
Any Free 2-level/hybrid/many-to-few Thread libraries? (Michael Thome)
Re: Linux Journal - worth or not? ("Steven J. Hill")
IP Masquerading seems flaky with 2.2.x kernels (Steve Mokris)
Re: JFS for linux? (David T. Blake)
----------------------------------------------------------------------------
From: Harald Arnesen <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.development.apps
Subject: Re: Linux Journal - worth or not?
Date: 30 Jul 1999 01:12:52 +0200
Reply-To: Harald Arnesen <[EMAIL PROTECTED]>
Bonn <[EMAIL PROTECTED]> writes:
> i started using linux for a while. is it worthwhile to subscribe the
> 'Linux Journal'?
> any comment?
Definitely - for all kinds of Linux users. It reminds me of Byte 15
years ago or so.
--
Harald Arnesen, Apall�kkveien 23 A, N-0956 Oslo, Norway
------------------------------
From: Gary Lawrence Murphy <[EMAIL PROTECTED]>
Subject: Call for Authors: SOHO System Admin Guide
Date: 29 Jul 1999 19:44:37 -0400
Reply-To: [EMAIL PROTECTED]
We need a lead author for a system administration guide for the
small office / home office (SOHO) level of user. Because the audience
will be people who are interested to learn how to install Linux as a
small office server, most of them may have very little unix sysadmin
experience --- this is your chance to stand on a pulpit and tell them
all how it *should* be done, as well as how to avoid all the pitfalls.
Because of the audience, we need someone with writing experience in a
journal, magazine or e'zine --- this is not a dummies guide, but it is
more of a SAG with added tutorials and unix essentials to get them up
to speed with both Linux and with managing the latest distributions
(eg linux-conf, gnome and kde utils &c)
Anyone interested can sign up at http://www.teledyn.com/authors.shtml
--
Gary Lawrence Murphy <[EMAIL PROTECTED]> TeleDynamics Communications Inc
Business Telecom Services : Internet Consulting : http://www.teledyn.com
Linux/GNU Education Group: http://www.egroups.com/group/linux-education/
"Computers are useless. They can only give you answers."(Pablo Picasso)
------------------------------
From: Karl Heyes <[EMAIL PROTECTED]>
Subject: Re: Writing shared libraries
Date: Thu, 29 Jul 1999 16:28:56 +0100
Kevin Woodward wrote:
>
> Thanks for clear that up. Are there any reasons for using static libs
> over shared libs. I've noticed that a lot of software comes with
> static libs...
>
static libs ease installations of some binary only packages, as the person
compiling the package can make sure that only the functions from the
static lib are actually used.
Useful for certain admin tools, where you suspect there might be a shared
library problem. For example a static command shell will enable an
administrator to get in and try to resolve any problems.
The only other reason is that share libs are position independant, so
there
might be slightly more time spent in execution, but it will be only about
1%.
The space/CPU/upgrade-ability spent using shared libs can often negate
any disadvantages with share libs.
karl.
------------------------------
From: Douglas Gilbert <[EMAIL PROTECTED]>
Crossposted-To:
linux.dev.c-programming,linux.dev.kernel,linux.dev.scsi,comp.os.linux.hardware
Subject: Re: Linux SCSI Performance Issues
Date: Thu, 29 Jul 1999 22:59:01 -0400
Dimi Shahbaz wrote:
>
> > Dave Platt wrote:
> >
> > Here are some quick results for 100MB sequential disk reads
> > on sg (2.2.10-ac10, advansys 940UW and 2 DCHS04U IBM disks):
> >
> > # time ./sg_dd512 if=/dev/sgb of=/dev/null count=200k
> > 204800+0 records in
> > 204800+0 records out
> > 0.00user 1.23system 0:10.44elapsed 11%CPU (0avgtext+0avgdata 0maxresident)k
> > 0inputs+0outputs (99major+27minor)pagefaults 0swaps
> >
> > # time ./sg_dd512 if=/dev/sgb of=/dev/null count=200k
> > 204800+0 records in
> > 204800+0 records out
> > 0.02user 1.75system 0:12.87elapsed 13%CPU (0avgtext+0avgdata 0maxresident)k
> > 0inputs+0outputs (99major+27minor)pagefaults 0swaps
> >
> > The first result was obtained without any other loading.
> > The second result was obtained when a similar command
> > was run on /dev/sga (which took 12.5 seconds elapsed).
> > When command queueing was used (i.e. sgq_dd512 command)
> > the coincident runs took 10.5 and 11.1 seconds each.
> >
> > Doug Gilbert
>
> I got some interesting results with the sg_dd512 program:
>
> time ./sg_dd512 if=/dev/sga of=/dev/null count=20k
> 20480+0 records in
> 20480+0 records out
> 0.00user 0.08system 0:03.52elapsed 2%CPU (0avgtext+0avgdata 0maxresident)k
> 0inputs+0outputs (88major+44minor)pagefaults 0swaps
>
> time ./sg_dd512 if=/dev/sga of=/dev/null count=20k &
> time ./sg_dd512 if=/dev/sgb of=/dev/null count=20k &
> 20480+0 records in
> 20480+0 records out
> 0.00user 0.11system 0:06.87elapsed 1%CPU (0avgtext+0avgdata 0maxresident)k
> 0inputs+0outputs (88major+44minor)pagefaults 0swaps
> 20480+0 records in
> 20480+0 records out
> 0.00user 0.12system 0:06.92elapsed 1%CPU (0avgtext+0avgdata 0maxresident)k
> 0inputs+0outputs (88major+44minor)pagefaults 0swaps
>
> time ./sg_dd512 if=/dev/sga of=/dev/null count=20k &
> time ./sg_dd512 if=/dev/sgb of=/dev/null count=20k &
> time ./sg_dd512 if=/dev/sgc of=/dev/null count=20k &
> 20480+0 records in
> 20480+0 records out
> 0.00user 0.12system 0:07.43elapsed 1%CPU (0avgtext+0avgdata 0maxresident)k
> 0inputs+0outputs (88major+44minor)pagefaults 0swaps
> 20480+0 records in
> 20480+0 records out
> 0.00user 0.16system 0:10.17elapsed 1%CPU (0avgtext+0avgdata 0maxresident)k
> 0inputs+0outputs (88major+44minor)pagefaults 0swaps
> 20480+0 records in
> 20480+0 records out
> 0.00user 0.15system 0:10.43elapsed 1%CPU (0avgtext+0avgdata 0maxresident)k
> 0inputs+0outputs (88major+44minor)pagefaults 0swaps
>
> Is this the sort of performance I'm supposed to get? The delay roughly
> doubles from 1 disk to 2 disks, and even more so from 1 disk to 3 disks (btw,
> the 2,3 disk runs were launched simultaneously to the background via a shell
> script). This was run on a 2.2.10 kernel, on 3 identical 9GB scsi disks.
Dimi,
No, these is not the results that are supposed to occur.
So now, when I run my own test (i.e. sg_dd512) and your test
(which you sent to me) I get good "parallel" behaviour
on my Advansys adapter. When you run the same tests on
your hardware you get bad "serialized" behaviour on your
???? adapter. Since NT has no problem with that adapter
then I am beginning to suspect the Linux driver for your
adapter (or its setup).
One thing I noticed about NT on my Advansys controller was
that it seemed to ignore the SCSI BIOS settings while the
Linux Advansys driver used them. So if 'disconnects' were
turned off in the SCSI BIOS you could get results like
those being seen by Dimi.
Note to Stefaan Eeckels: This doesn't seem like a pthread's
issue since my test (i.e. 2 'sg_dd512's) involves separate
processes.
Doug Gilbert
------------------------------
From: jievis <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.development.apps,comp.os.linux.misc
Subject: The Lex & Yacc source code for C++
Date: Fri, 30 Jul 1999 02:30:30 GMT
Hi,
Who can tell me where can found the source codes of lex and yacc for
c++. it is better if the grammers for preprocesser is included.
Thank in advance.
Jievis
================== Posted via SearchLinux ==================
http://www.searchlinux.com
------------------------------
From: Andreas Jaeger <[EMAIL PROTECTED]>
Subject: Re: How to find Network interface index?
Date: 29 Jul 1999 09:58:17 +0200
>>>>> Jung-gyun Ahn writes:
> PF_PACKET socket need Interface number to send packet.
> Past packet socket required a network interface name like "eth0".
> It use dev_get() call to find device structure.
> But new packet socket need a Interface number. and use
> dev_get_by_ifindex().
> How to find ifindex in device structure in user (NOT KERNEL) mode?
>From <net/if.h> in glibc 2.1.x:
/* Convert an interface name to an index, and vice versa. */
extern unsigned int if_nametoindex __P ((__const char *__ifname));
extern char *if_indextoname __P ((unsigned int __ifindex, char *__ifname));
/* Return a list of all interfaces and their indices. */
struct if_nameindex
{
unsigned int if_index; /* 1, 2, ... */
char *if_name; /* null terminated name: "eth0", ... */
};
extern struct if_nameindex *if_nameindex __P ((void));
/* Free the data returned from if_nameindex. */
extern void if_freenameindex __P ((struct if_nameindex *__ptr));
Otherwise use directly SIOCGIFCONF.
Andreas
--
Andreas Jaeger [EMAIL PROTECTED] [EMAIL PROTECTED]
for pgp-key finger [EMAIL PROTECTED]
------------------------------
From: Adam Przybyla <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.development.apps,comp.os.linux.misc
Subject: Re: The Lex & Yacc source code for C++
Date: 30 Jul 1999 07:04:01 GMT
In comp.os.linux.development.apps jievis <[EMAIL PROTECTED]> wrote:
> Hi,
> Who can tell me where can found the source codes of lex and yacc for
> c++. it is better if the grammers for preprocesser is included.
> Thank in advance.
... new bison 1.27 sopports C++ ;-) Regards
Adam Przybyla
------------------------------
Date: Fri, 30 Jul 1999 10:26:09 +0200
From: Stefan Opperskalski <[EMAIL PROTECTED]>
Subject: JFS for linux?
Hi,
does anybody know, where I can get informations on the JFS-project for
linux?
I somewhere read, that a port from SUN will take place...
Thanks,
Stefan O.
------------------------------
From: "Chris Liddell" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.development.apps
Subject: Re: Linux Journal - worth or not?
Date: Fri, 30 Jul 1999 08:55:18 +0100
After subscribing to LJ for over two years, my feeling is it's definitely
worth while. As observed below, there's not *always* alot in it, I usually
find that there's two articles in every issue I really want to read, then
every few months there's a issue where I want to read everything in it.
It is excellent value for money!
Chris Liddell
Christopher Browne wrote in message ...
>On Fri, 30 Jul 1999 01:11:17 +0800, Bonn <[EMAIL PROTECTED]> wrote:
>>i started using linux for a while. is it worthwhile to subscribe the
>>'Linux Journal'?
>>any comment?
>
>I've been subscribing for about 3 years now.
>
>It's not always *tremendously* worthwhile, but there are almost always
>some interesting articles, and usually something illuminating.
>--
>"Few people are capable of expressing with equanimity opinions which
>differ from the prejudices of their social environment. Most people are
>even incapable of forming such opinions." (Albert Einstein)
>[EMAIL PROTECTED] <http://www.ntlug.org/~cbbrowne/linux.html>
------------------------------
From: [EMAIL PROTECTED] (Matthias Maisenbacher)
Subject: Re: Compiling Linux 1.0
Date: 29 Jul 1999 11:40:25 GMT
Reply-To: [EMAIL PROTECTED]
I asked:
>Is there any receipe for what i'm doing wrong or do i have
>to repair all errors one by one?
Many others answered. Thank you.
But in the meantime I decided to install an original 1.0(.9) Linux box.
This give me one more Linux installation :-)
And a much easier compile of this old kernel.
Thank you to all who contributed
Matthias
------------------------------
Crossposted-To: comp.os.linux.development.apps,comp.os.linux.misc
Subject: Re: The Lex & Yacc source code for C++
From: "Noah Roberts (jik-)" <[EMAIL PROTECTED]>
Date: 30 Jul 1999 01:45:01 -0700
Adam Przybyla <[EMAIL PROTECTED]> writes:
> In comp.os.linux.development.apps jievis <[EMAIL PROTECTED]> wrote:
> > Hi,
> > Who can tell me where can found the source codes of lex and yacc for
> > c++. it is better if the grammers for preprocesser is included.
> > Thank in advance.
> ... new bison 1.27 sopports C++ ;-) Regards
> Adam Przybyla
I think he wants the lex and yacc files to parse C++...not the lex and
yacc that support C++ output.
------------------------------
From: [EMAIL PROTECTED] (Dr H. T. Leung)
Subject: Re: glibc 2.1.1 clock() problem (returning -ve number?)
Date: 30 Jul 1999 11:30:51 GMT
So this means my libc5 binary will also wrap around eventually (it would be
almost 8 months if I let my computer stay up for that long)? But I was almost
following the info page (even the little snipplet code there), I think the info
page is a bit misleading... I mean, timing CPU usage for 40 minutes (it is only
1/2 of 71min if signed) is not much use, really...
By the way, I found that the libc6 math library (for my purpose) is about 5%
faster than the libc5 math library, but it starts up slower (the initial stage
took longer).
In article <[EMAIL PROTECTED]>, Andreas Jaeger
<[EMAIL PROTECTED]> writes:
|> >>>>> H T Leung writes:
|>
|> > I have a big (long-running) program that output CPU time usage
|> > continously with this routine; compile with libc 5 it gives the
|> > correct output, but when compiled and linked against glibc 2.1.1,
|> > [..]
|> > I am sure it is some silly thing I do, but I can't seem to figure
|> > out why it does that. I think under libc 5 CLOCKS_PER_SEC is 100,
|> > while in Glibc 2 it is much bigger(1e6), but it still doesn't
|> > account for the change.
|>
|> Do a bit of mathematics. clock_t is 32 bit on ix86, CLOCKS_PER_SEC is
|> required by POSIX to be 1e6 this leads to a wrap around of clock_t
|> after 71 minutes. Everything is working as expected - only your
|> expectations aren't right;-)
|>
|> There are better functions for longer running processes like
|> getrusage.
|>
|> Andreas
|> --
|> Andreas Jaeger [EMAIL PROTECTED] [EMAIL PROTECTED]
|> for pgp-key finger [EMAIL PROTECTED]
--
--------------------------------------------------
"What you don't care cannot hurt you." Chap. 7a, AMS-NS
------------------------------
From: [EMAIL PROTECTED] (J.H.M. Dassen (Ray))
Subject: Re: JFS for linux?
Date: 30 Jul 1999 11:36:49 GMT
Stefan Opperskalski <[EMAIL PROTECTED]> wrote:
>does anybody know, where I can get informations on the JFS-project for
>linux?
I'll assume JFS == journalling filesystem.
Have a look at http://linuxtoday.com/stories/5556.html for a large list of
filesystem-related projects.
>I somewhere read, that a port from SUN will take place...
SGI has announced it will port its XFS to Linux (see e.g.
http://linuxtoday.com/stories/6149.html).
HTH,
Ray
--
PATRIOTISM A great British writer once said that if he had to choose
between betraying his country and betraying a friend he hoped he would
have the decency to betray his country.
- The Hipcrime Vocab by Chad C. Mulligan
------------------------------
From: Michael Thome <[EMAIL PROTECTED]>
Crossposted-To: comp.programming.threads
Subject: Any Free 2-level/hybrid/many-to-few Thread libraries?
Date: Fri, 30 Jul 1999 13:13:12 GMT
Are there any freely-available (GPL, PD, whatever) thread libraries
available (esp. for Linux) which implement a mixed "many user-space
threads to a few kernel-space threads" model?
User-space threads do not benefit from SMP, at least not directly,
whereas Kernel-space threads are overly expensive and seriously
resource-limited. Yes, I'm familiar with the standard arguments that
using (say) hundreds of threads is bad design... there are, however, at
least two legit (IMHO) situations where support of many threads is
required:
1. Threads as a language construct (e.g. in Java, etc). Where threads
are integral to the language (as much as closures, continuations,
coprocesses, etc) there cannot be low resource limits on threads without
damaging the integrity of the the language.
2. Development of peer-to-peer distributed applications. Even though a
single peer may require a small number of threads, development of the
full application may require tens or even hundreds of peers. Even if
most peers have small resource requirements (and should be capable of
running on a single machine), you can easily fill up a linux process
table if you are using native threads.
Solaris seems to have an interesting 2-level thread implementation - I'd
love to see a similar one for linux...
thanks,
-mik
Michael Thome
[EMAIL PROTECTED]
------------------------------
From: "Steven J. Hill" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.development.apps
Subject: Re: Linux Journal - worth or not?
Date: Fri, 30 Jul 1999 09:16:17 -0500
I would tend to agree with Mr. Rempt's initial opinion that the Linux
Journal is becoming more commercially oriented. I just recently started
to subscribe to Linux Magazine which IMHO is has good technical content
and is not so full of advertisements. I have the first three issues and
have read them all cover to cover. Here is the URL:
http://www.linux-mag.com/
-Steve
======================================================
Steven J. Hill
Diversified Software Industries, Inc.
625 South Gilbert Suite 2
Iowa City, IA 52240
E-mail: [EMAIL PROTECTED] URL: http://www.dsi-inc.net
Phone: (319) 337-2668 ext. 105
------------------------------
From: Steve Mokris <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.networking
Subject: IP Masquerading seems flaky with 2.2.x kernels
Date: Fri, 30 Jul 1999 09:55:27 -0400
I have recently upgraded several network gateway machines to various
versions of the 2.2.x kernel (where x=6, 7, and 10). I was previously
running 2.0.34 which I compiled to include support for IP Masquerading,
and this setup worked very well. I used ipfwadmin to initialize the
masquerading.
I then converted to the various 2.2.x kernels, followed the howto, and
converted all of the ipfwadmin usage to ipchains and made sure to enable
ip forwarding. Here is the current, relevant portion of my
/etc/rc.d/rc.inet:
---
# 42.42.42.42 is a dummy ip to initialize eth0 to before it gets
# the real ip from dhcp. it did not seem to initialize properly
# otherwise.
/sbin/ifconfig eth0 42.42.42.42 up
/sbin/rrdhcpcd
/sbin/rrlogind
/sbin/ifconfig eth0:0 10.0.7.1 up
/sbin/route add -net 10.0.7.0 netmask 255.255.255.0 eth0:0
/sbin/ipchains -P forward DENY
/sbin/ipchains -A forward -s 10.0.7.0/24 -i eth0 -j MASQ
echo 1 > /proc/sys/net/ipv4/ip_forward
/sbin/modprobe ip_masq_ftp
/sbin/modprobe ip_masq_user
---
I have had varying success with external access using machines on the
local network (10.0.7.0/24). Windows machines (arrgh) and Unix
workstations (two SGI Indys, an HP workstation, an AlphaStation, and a
Sun) have no trouble accessing the Internet through the masquerading
gateway, though all of the Linux boxen have very flaky access.
After rebooting, I can (for example) ping netscape.com seemingly
indefinitely. I then try to ping freshmeat.net, and get at most four
ping responses. Following that, I am unable to access any other
external servers. The same thing happens with any set of two external
servers.
I have read some FAQ's and newsgroups, and tried two things:
* I applied the patch from http://juanjox.linuxhq.com/ and
recompiled the kernel. This slightly decreased the flakiness
(I'm sorry I need to refer to it in such qualitative terms,
though the problem seems very transient and difficult to test)
though it is still very unreliable.
* The MTU on both the gateway and one internal machine I tried
was set to 1500. I tried setting this on all devices on both
machines to 296 as per one FAQ. This did not solve the problem
and only seemed to significantly slow down external transfers.
Has anyone else encountered this problem? I have followed the howto and
completely started over on three instances, with three kernel versions
on two separate computers, and encountered precisely the same problem,
which leads me to believe it may be a kernel bug (or configuration
braindamage on my part). Any suggestions would be appreciated.
Thank you,
Steve
------------------------------
From: [EMAIL PROTECTED] (David T. Blake)
Subject: Re: JFS for linux?
Date: 30 Jul 1999 13:53:53 GMT
Reply-To: [EMAIL PROTECTED]
Stefan Opperskalski <[EMAIL PROTECTED]> wrote:
> does anybody know, where I can get informations on the
> JFS-project for linux?
Stephen Tweedie is working on a beta of ext3 that will be journalled
and compatible with the VFS in the current devel series. He
has not publicly released any of it yet, and is behind his
previously stated schedule. Such as it goes.
> I somewhere read, that a port from SUN will take place...
SGI is porting XFS. Although if you say "show me the code" you
won't see any. They haven't even released their putative license
for the release.
--
Dave Blake
[EMAIL PROTECTED]
------------------------------
** 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
******************************
