Linux-Development-Sys Digest #836, Volume #6 Wed, 16 Jun 99 08:14:09 EDT
Contents:
Re: TAOs: Much to do about nothing? (The Ghost In The Machine)
Re: Any Journaling FS development? (Bob Tennent)
Re: Pinning a thread to a processor (bill davidsen)
adduser + passwd command through a C script ? Help ([EMAIL PROTECTED])
Re: What are the differences between mySQL and mSQL? (david parsons)
Re: Current process count (Josef =?iso-8859-1?Q?M=F6llers?=)
Re: Using pentiums debug registers under linux (James Cownie)
Re: Kernel Hang ( linux 2.2.x x<=10 ) (Stepan Kasal)
Re: TAO: the ultimate OS (void)
core dump problem? (Kong Tae Young)
Problem about time slice of process/threads ("Cligon")
Re: SIGALRM and serial i/o (Andreas Schwab)
Re: Current process count (Andreas Schwab)
Protocol features for Linux?
Re: TAOs: Much to do about nothing? (Donal K. Fellows)
Re: core dump problem? (Josef =?iso-8859-1?Q?M=F6llers?=)
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (The Ghost In The Machine)
Crossposted-To: alt.os.linux,comp.os.linux.advocacy,comp.os.misc,comp.unix.advocacy
Subject: Re: TAOs: Much to do about nothing?
Date: Tue, 15 Jun 1999 18:25:43 GMT
On 15 Jun 1999 09:00:11 -0500, Brian McGroarty <[EMAIL PROTECTED]> wrote:
>[EMAIL PROTECTED] (Sam Holden) writes:
>
>>On 15 Jun 1999 10:33:09 GMT, Gordon Scott <[EMAIL PROTECTED]> wrote:
>>>Technolord ([EMAIL PROTECTED]) wrote:
>>>: He just requires to know what can and cannot
>>>: be done (placing a bed in a 1.5mx1.5m room isn't possible) and then
>>>
>>>Off topic I guess, but that is possible if you think a little laterally,
>>>The diagonal in that room is 2.1m and there's still (just!) room for a
>>>door, window and clothes locker. It'll be a little cozy.
>>>
>>>I'm ignoring children's rooms. (Too easy ;-)
>
>>And the fact that beds are not 1 dimensional objects...
>
>
>/me stands the bed on end...
>
>Nor are rooms. ;)
Yeah, but unless the bed has straps on it (or one is in a
space station with zero gravity), how is one going to get
*into* the bed and *sleep* in it?
[.sigsnip]
----
[EMAIL PROTECTED] -- unlike bras, most beds are strapless... :-)
------------------------------
From: [EMAIL PROTECTED] (Bob Tennent)
Subject: Re: Any Journaling FS development?
Date: 15 Jun 1999 20:17:21 GMT
Reply-To: rdt(a)cs.queensu.ca
On 10 Jun 1999 17:19:38 -0500, [EMAIL PROTECTED] wrote:
>In article <[EMAIL PROTECTED]>,
>David T. Blake <[EMAIL PROTECTED]> wrote:
>>[EMAIL PROTECTED] writes:
>
>>According to SGI, it is so that their lawyers can look into
>>it and see how 'freely' they will be allowed to license
>>XFS
>
>hoo boy. What lawyer wants to be the first to give that
>kind of legitimacy to GPL? I'm not holding my breath anymore.
I think you've missed the point. The lawyers are presumably
checking that freeing the code doesn't violate any *existing*
licensing restrictions. The same problem arose for Netscape.
Bob T.
------------------------------
From: [EMAIL PROTECTED] (bill davidsen)
Subject: Re: Pinning a thread to a processor
Date: 15 Jun 1999 23:29:11 GMT
In article <7jr4p3$c8r$[EMAIL PROTECTED]>,
Karsten Scholtyssik <[EMAIL PROTECTED]> wrote:
|
| bill davidsen <[EMAIL PROTECTED]> schrieb in im Newsbeitrag:
| 7jpdgk$d90$[EMAIL PROTECTED]
| > Out of curiousity, what do you think this binding with do for you?
| It prevents the process from beeing scheduled on a different CPU
| which increases the cache locality. It reduces the situations where
| the process has to start with a cold cache.
But Linux already preferentially uses the previous processor. You
*don't* want to always use the same processor, since some other process
may have been run on it and you get two performance hits, one from
waiting for the processor to be free and one from reloading the cache
anyway after the other process uses it.
I complain about kernel choices on a regular basis, but after tuning the
afinity constant a bit, I find that in a real situation the system runs
better with a slight bias to the previous processor than a lot.
| > Also, does this NT call bind the process to the processor or the
| > processor to the process?
| It binds the process to the processor so that the process is only
| scheduled on the given Processor(s).
You can get that behaviour in Linux with a simple tuning of the affinity
constant, but if the number of processes is greater than the number of
processors performance will go down.
================ possible patch ================
If someone wanted to get really brave they could add a bind system call
which bound the L2 cache to a given process, and disabled the L2 when
the kernel or another user process was using it. That would avoid *most*
of the cache flush for the chosen process, without introducing a huge
penalty to the other processes.
The bottom line is that a cache flush just doesn't hurt that much, at
least unless you have a program which fits all in cache. Even then you
lose microsec, so waiting longer than that for the "right" processor is
an overall loss.
--
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]
Subject: adduser + passwd command through a C script ? Help
Date: Wed, 16 Jun 1999 04:01:29 GMT
I would like to create a HTML form for visitors to create a user accoutn
on my Linux Redhat 6 (shadow passwords) system.
Anyone know a good C cgi scirpt to do this ?
Thank you
[EMAIL PROTECTED]
Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
------------------------------
From: o r c @ p e l l . p o r t l a n d . o r . u s (david parsons)
Crossposted-To:
comp.os.linux.development.apps,comp.os.linux.misc,comp.os.linux.setup,comp.lang.java.databases
Subject: Re: What are the differences between mySQL and mSQL?
Date: 15 Jun 1999 21:40:28 -0700
In article <[EMAIL PROTECTED]>,
Don Baccus <[EMAIL PROTECTED]> wrote:
>In article <7k1vru$[EMAIL PROTECTED]>,
>david parsons <o r c @ p e l l . p o r t l a n d . o r . u s> wrote:
>>In article <[EMAIL PROTECTED]>,
>>Don Baccus <[EMAIL PROTECTED]> wrote:
>
>>>Not necessarily true. If you have several related tables
>>>that need to be logically updated at once, the atomicity of
>>>the transactional model is, well, useful if there's a crash
>>>while records are being inserted or updated.
>
>> Denormalize, denormalize, denormalize.
>> Yeah, you might bloat your rdb by a factor of 10 to do this,
>> but disk and core is getting cheap these days.
>
>Denormalization is something sane folks do to increase performance,
>not to avoid the need for a real database engine.
You'd be surprised; once you get away from the database
pundits and RDB consultants (both types favor correctness
when it means that more RDB software will be purchased from
their corporate masters), about every permutation you can
think of will be used in the real world.
You don't want to think about what people will do to increase
performance. Running a hot database on a server that doesn't
do transactions is pretty minor stuff compared to some of the
things that can be done.
>If you've denormalized checking balances into thirty tables,
>don't have atomicity, and an update dies half-way through
>updating these tables, which balance do you trust?
Then you've (a) not denormalized properly and (b) an idiot for not
using a transactional database for a problem that won't work
without transactions.
I'd consider a checking balance to be a report field, and just use
a report generator to generate it out of the ledger. You could
probably do it in a nontransactional system; you'd have all the
credits and debits be atomic inserts into the ledger, run a
periodic job to build the balance report, then join against the
balance report whenever you need the checking balance.
You could. I wouldn't, unless I had a six-month retainer in my
pocket and a notarised amnesty protecting me from legal action
when some clever hacker kites the first $100,000 check.
>Why bother? Why not use a transaction-based db in the first
>place?
Dunno. Possibly for performance reasons, possibly for religious
reasons (cf: Sybase not being Free(tm)), possibly because of
sheer cussedness.
____
david parsons \bi/ Working for Sybase, like having a police record,
\/ leaves a stain on your reputation forever.
------------------------------
From: Josef =?iso-8859-1?Q?M=F6llers?= <[EMAIL PROTECTED]>
Subject: Re: Current process count
Date: Wed, 16 Jun 1999 08:42:32 +0200
Sureshkumar Kaliannan wrote:
> =
> Hi,
> =
> I am new to kernel programming and I need the following information.
> =
> I am writing a module that requires the current process count( the numb=
er
> of active processess including ready,running,sleeping processes).
> =
> Does the kernel maintain it in some variable that is exported???
> =
> Currently, I calculate it as follows...
> =
> int count =3D 0;
> struct task_struct *p;
> for (p =3D current ; (p =3D p->next_task) !=3D current ; )
> ++count;
> =
> I couldn't use
> #define for_each_task(p) \
> for (p =3D &init_task ; (p =3D p->next_task) !=3D &init_task ; =
)
> since "init_task" is not exported by the kernel.
> =
> Do I have to rebuild the kernel to export it?? (If so how do I do it??)=
> I am using the kernel that came with RH 5.2.
> =
> Thanks for any help
> Suresh
Read /proc/loadavg. The 4th field has the number of processes after the
slash.
-- =
PS Die hier dargestellte Meinung ist die persoenliche Meinung des
Autors!
PS This article reflects the autor=B4s personal views only!
------------------------------
Date: Wed, 16 Jun 1999 09:25:28 +0100
From: James Cownie <[EMAIL PROTECTED]>
Crossposted-To: comp.sys.intel,de.comp.lang.assembler.x86,alt.os.linux,alt.linux
Subject: Re: Using pentiums debug registers under linux
Markus Jochim wrote:
>
> Hello everybody,
>
> I have some questions concerning the debug-registers of the Intel
> Pentium processors and how to use them under linux.
>
> (It is not helpful to give me any hints about which program could
> be used for debug reasons because I do not want to use the
> debug registers in order to debug, but I want to use them in order
> to (for example) manipulate some register values at a specific point
> in an application in order to test the fault-tolerance of such an
> application)
>
> What I want i the following:
>
> An application program must run under linux and it must be able
> to set breakpoints for this application by using the debug registers
> DR0 - DR3 and DR6, DR7.
>
You suggest all kinds of complexity. However, the ptrace calls allow
you to set the debug registers (via the Uarea access calls), and you
can then get the reports back via ptrace too.
Take a look at the code in arch/i386/kernel/ptrace.c for the kernel
side, and the code in gdb for the user end.
-- Jim
James Cownie <[EMAIL PROTECTED]>
Etnus, Inc. +44 117 9071438
http://www.etnus.com
------------------------------
From: [EMAIL PROTECTED] (Stepan Kasal)
Crossposted-To: linux.dev.kernel
Subject: Re: Kernel Hang ( linux 2.2.x x<=10 )
Date: 16 Jun 1999 09:16:51 GMT
On Tue, 15 Jun 1999 17:32:56 +0000, Charles-Edouard Ruault wrote:
> It's an AMD K6-2 400 on an epox motherboard, 128MB RAM , IBM IDE HDD ,
Hi,
if you experience hangups, it may be coused by the memory. Do you have
another chips handy? Is your configuration OK according what motherboard
docs say? I've heard about cases where the same machine worked OK with MS
things but hanged under Linux.
Maybe it helps, Stepan
------------------------------
From: [EMAIL PROTECTED] (void)
Crossposted-To: alt.os.linux,comp.os.linux.advocacy,comp.os.misc,comp.unix.advocacy
Subject: Re: TAO: the ultimate OS
Date: 16 Jun 1999 02:58:03 GMT
On Tue, 15 Jun 1999 15:43:18 GMT, Bill Vermillion <[EMAIL PROTECTED]>
wrote:
>
>Heck - I was taught to clap with one hand years and years ago.
Well then, smart-ass, what is the sound of half a hand clapping?
--
Ben
"The world is conspiring in your favor." -- de la Vega
------------------------------
From: Kong Tae Young <[EMAIL PROTECTED]>
Subject: core dump problem?
Date: Wed, 16 Jun 1999 18:18:01 +0900
==============AF72492C30B71A6D42FF88C2
Content-Type: text/plain; charset=EUC-KR
Content-Transfer-Encoding: 7bit
I have just upgraded to RedHat 6.0(kernel 2.2.9 ), and my program ,
which
generates segmentation fault error intentionally, doesn't generate core
dump file
when segmentation fault error happens.
I want to generate core dump file when a process goes wrong.
Especially core dump file's good tool for debugging.
It worked without any problem before upgrading.
I tried "ulimit -c unlimited" in bash shell(2.0.3) and didn't work.
When I changed my shell from bash to csh, following comment appeared.
coredumpsize : can't set limit.
What's wrong? and What should I do? Help me.
--
Below code generated core dump, but now it doesn't
main
{
char *name = "";
printf("name is %s", name[5]);
}
===========================
Kong Tae Young
Dep of Computer Engineering
Dong-A Univ, South Korea
[EMAIL PROTECTED]
============================
==============AF72492C30B71A6D42FF88C2
Content-Type: text/html; charset=EUC-KR
Content-Transfer-Encoding: 7bit
<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
I have just upgraded to RedHat 6.0(kernel 2.2.9 ), and my program
, which
<br>generates segmentation fault error intentionally, doesn't generate
core dump file
<br>when segmentation fault error happens.
<p>I want to generate core dump file when a process goes wrong.
<br>Especially core dump file's good tool for debugging.
<p>It worked without any problem before upgrading.
<p>I tried "ulimit -c unlimited" in bash shell(2.0.3) and didn't work.
<br>When I changed my shell from bash to csh, following comment appeared.
<p>coredumpsize : can't set limit.
<p>What's wrong? and What should I do? Help me.
<pre>--</pre>
<pre>Below code generated core dump, but now it doesn't</pre>
<pre>main</pre>
<pre>{</pre>
<pre>char *name = "";</pre>
<pre>printf("name is %s", name[5]);</pre>
<pre>}</pre>
<pre> </pre>
<pre>
===========================
Kong Tae Young
Dep of Computer Engineering
Dong-A Univ, South Korea
[EMAIL PROTECTED]
============================</pre>
</html>
==============AF72492C30B71A6D42FF88C2==
------------------------------
From: "Cligon" <[EMAIL PROTECTED]>
Subject: Problem about time slice of process/threads
Date: Wed, 16 Jun 1999 17:23:13 +0800
Hello,
I'm new in unix programming.
The document about scheduling I've found
only provide scheduling policy and priority of process/threads.
Recently I've met a problem:
if I want to customize the time slice,
is there any API provide this function?
Thanks a lot.
------------------------------
From: Andreas Schwab <[EMAIL PROTECTED]>
Subject: Re: SIGALRM and serial i/o
Date: 16 Jun 1999 11:44:03 +0200
<[EMAIL PROTECTED]> writes:
|> I'm posting the minimal source code for my serial port reading program.
|> The problem didn't seem to be the SIGALRM code in itself. It seems that
|> the serial port settings get reset or overwritten if I don't keep the
|> termios structure in memory.
You are not correctly initializing the memory pointed to by newtio. You
must copy the old state into it before you can modify it in a meaningfull
way.
|> newtio=malloc(sizeof(struct termios));
|> if( newtio==NULL ){ perror("Malloc"); exit (-1);}
|>
|> fd = open("/dev/ttyS0", O_RDWR | O_NOCTTY );
|> if( fd < 0 ){ perror("open"); return -1;}
|>
|> ret= (tcgetattr(fd, &oldtio));
|> if(ret < 0){ perror("tcgetattr"); exit(-1);}
*newtio = oldtio;
|> newtio->c_cflag |= CLOCAL | CREAD;
|> newtio->c_cflag |= CS8;
|>
|> newtio->c_iflag |= INPCK | ISTRIP;
|> newtio->c_iflag &= ~IUCLC;
|>
|> newtio->c_lflag |= ICANON;
--
Andreas Schwab "And now for something
[EMAIL PROTECTED] completely different"
[EMAIL PROTECTED]
------------------------------
From: Andreas Schwab <[EMAIL PROTECTED]>
Subject: Re: Current process count
Date: 16 Jun 1999 11:46:05 +0200
Sureshkumar Kaliannan <[EMAIL PROTECTED]> writes:
|> Hi,
|>
|> I am new to kernel programming and I need the following information.
|>
|> I am writing a module that requires the current process count( the number
|> of active processess including ready,running,sleeping processes).
Are you sure that it *really* requires it?
--
Andreas Schwab "And now for something
[EMAIL PROTECTED] completely different"
[EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] ()
Subject: Protocol features for Linux?
Date: 16 Jun 1999 10:19:11 GMT
Reply-To: [EMAIL PROTECTED]
I am currently investigating the availability and functionality
of a number of protocols for Linux. Please let me know if they are
available (in any form, even outside the standard linux kernel),
and if they aren't in the standard kernel then please tell me where
I can find them (if you can).
Transport Protocols
* IPX/SPX (Does it support the services SAP, RIP and Local
Bindery Database?)
* Netware/IP
* Pure IP (Netware)
* NetBEUI (If not is there still any support for llc2?)
* TCP/IP (Does it support ipv6 fully, I can't find much updated info
here?)
Configuration Protocols
* SNMP (Need atleast version 1.0 supporterted and does it have TRAP
support?)
* SLP (Does it have Directory Agent Support?)
Access/Application Protocols
* AFP (Linux is supposed to have 2.0 support, but I need 2.2 support).
* SMB (I need support for SMB over TCP/IP, NetBEUI and IPX/SPX)
* NFS (Does it have Locking Mechanism (NLM)?)
* NetWare (Need support for up to v5.0).
File Systems
* ISO9660 (Rock Ridge, Joliet extensions and High Sierra?)
* UDF
If anyone knows where i can find info on these subjects that is
also very much appreciated.
Thanks in advance for all your help, and please cc to the mailaddy
above (and remove the trailing ".fake").
=======================
Niclas Anderberg
Axis Communications AB
Lund, Sweden
email: [EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Donal K. Fellows)
Crossposted-To: alt.os.linux,comp.os.linux.advocacy,comp.os.misc,comp.unix.advocacy
Subject: Re: TAOs: Much to do about nothing?
Date: 16 Jun 1999 10:56:15 GMT
In article <[EMAIL PROTECTED]>,
The Ghost In The Machine <[EMAIL PROTECTED]> wrote:
> Brian McGroarty <[EMAIL PROTECTED]> wrote:
>> /me stands the bed on end...
>> Nor are rooms. ;)
>
> Yeah, but unless the bed has straps on it (or one is in a space
> station with zero gravity), how is one going to get *into* the bed
> and *sleep* in it?
Nobody said anything about *using* the bed! You're changing the
specification at runtime...
Donal.
--
Donal K. Fellows http://www.cs.man.ac.uk/~fellowsd/ [EMAIL PROTECTED]
-- The small advantage of not having California being part of my country would
be overweighed by having California as a heavily-armed rabid weasel on our
borders. -- David Parsons <o r c @ p e l l . p o r t l a n d . o r . u s>
------------------------------
From: Josef =?iso-8859-1?Q?M=F6llers?= <[EMAIL PROTECTED]>
Subject: Re: core dump problem?
Date: Wed, 16 Jun 1999 12:51:40 +0200
Kong Tae Young wrote:
> =
> I have just upgraded to RedHat 6.0(kernel 2.2.9 ), and my program ,
> which
> generates segmentation fault error intentionally, doesn't generate
> core dump file
> when segmentation fault error happens.
> =
> I want to generate core dump file when a process goes wrong.
> Especially core dump file's good tool for debugging.
> =
> It worked without any problem before upgrading.
> =
> I tried "ulimit -c unlimited" in bash shell(2.0.3) and didn't work.
> When I changed my shell from bash to csh, following comment appeared.
> =
> coredumpsize : can't set limit.
> =
> What's wrong? and What should I do? Help me.
> =
> --
> =
> Below code generated core dump, but now it doesn't
> =
> main
> =
> {
> =
> char *name =3D "";
> =
> printf("name is %s", name[5]);
> =
> }
This is very implementation dependent.
The empty string ("") is allocated somewhere in the virtual address
space of the process.
Anything may follow it, so accessing name[5] may or may not fail in the
first place.
If it does not fail (because name[5] lies within the virtual address
space of the process), printf may fail because the value passed (which
is a character casted into a pointer) may be invalid.
Probebly the old libc core dumped when printf tried to print the string
while the new one is more lenient.
If you definitely need to core dump, you might consider using abort(),
however the manual doesn't specify that it dumps core, or write a value
to virtual address 0, whcih, again, is implementation dependent.
Josef
-- =
PS Die hier dargestellte Meinung ist die persoenliche Meinung des
Autors!
PS This article reflects the autor=B4s personal views only!
------------------------------
** 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
******************************
