Linux-Development-Sys Digest #621, Volume #8 Mon, 9 Apr 01 14:13:12 EDT
Contents:
Re: setitimer() problem on linux (Kasper Dupont)
Re: Kernel mode to User Mode (Kasper Dupont)
Re: usleep() is unreliable when sleeping for less then 10000 micro (Rolf Magnus)
how to modify dynamic lib ? (Markus Fischer)
Suppressing Redhat bootup output (Paul Haley)
ALIGN in entry.S (Mikael Chambon)
Re: Suppressing Redhat bootup output ("Chris West")
Re: Suppressing Redhat bootup output (Kasper Dupont)
Re: ALIGN in entry.S (Kasper Dupont)
how to modify dynamic lib ? (Markus Fischer)
Re: process becomes [process] (Kasper Dupont)
Re: ALIGN in entry.S (Mikael Chambon)
Re: usleep() is unreliable when sleeping for less then 10000 micro (Pasztor Szilard)
Re: Need your recommendation for a full-featured text editor (Greg Copeland)
DEV TCP/IP C++ ("Dirk Riebesell")
Re: setitimer() problem on linux (Matthew Gatto)
----------------------------------------------------------------------------
From: Kasper Dupont <[EMAIL PROTECTED]>
Subject: Re: setitimer() problem on linux
Date: Mon, 09 Apr 2001 13:26:54 +0000
Matthew Gatto wrote:
>
> I wrote a simple process scheduling shell for school, that I developed
> on a SUN-OS Unix server. The program works fine there, but on my home
> linux computer, it is hanging during a setitimer() call.
>
> After selecting a child process to give a timeslice to, it sends it
> the start signal then does the following code (that should pause a
> short time, and then wake up and continue executing the process
> scheduler).
>
> the struct itimerval was initialized as follows at the begining of the
> program:
> struct itimerval new;
> new.it_interval.tv_usec = 0;
> new.it_interval.tv_sec = 0;
> new.it_value.tv_usec = 600;
> new.it_value.tv_sec = 0;
>
> and the sigaction initialised as follows:
> struct sigaction action;
> sigfillset(&(action.sa_mask));
> action.sa_handler=do_nothing; /* do_nothing(), is as the name
> indicates, an empty function */
> sigaction(SIGALRM, &action, NULL);
>
> and during the scheduling loop the following code gets executed:
>
> /* start the timer */
>
> if (DEBUG ==1)
> { printf("about to set itimer\n"); }
> setitimer(ITIMER_REAL, &new, NULL);
>
> if (DEBUG == 1)
> { printf("itimer set, about to pause()!"); }
> pause();
>
> it hangs consistently on the first or second time (on linux at least)
> through this code block. it will print out
>
> $
> about to set itimer
> itimer set, about to pause()!
> about to set itimer
>
> and then hang indefinately.. not responding to any signals
> (SIGINT,SIGTSTP,SIGCONT,etc). if I do a ps -el on another terminal, it
> shows the program's WCHAN as doing a "pause". does anyone know why
> this could be? or what i could be doing something wrong? i am truly
> clueless.
>
> I posted the code here:
> http://www.ductape.net/~gatto/scheduler/
> The timer call that hangs is in
> http://www.ductape.net/~gatto/scheduler/main.c
You forgot to initialize the sa_flags.
I added this line and then it worked for me:
action.sa_flags=0;
--
Kasper Dupont
------------------------------
From: Kasper Dupont <[EMAIL PROTECTED]>
Subject: Re: Kernel mode to User Mode
Date: Mon, 09 Apr 2001 13:40:33 +0000
Mikael Chambon wrote:
>
> Hi,
>
> Could someone explain me clearly what is the difference between the
> kernel mode and user mode ??
>
> In the book "Understanding the linux kernel p 252", I road that
> the kernel checks if a signal has arrived when switching from user
> mode to kernel mode.
It is not when switching from user mode to kernel mode,
it is when switching from kernel mode back to user mode.
>
> Can someone give me the .c file where this is done ?? I would like
> to add something..
It is not done in a .c file, it is done in assembler
code.
arch/i386/kernel/entry.S
The do_signal() function is in
arch/i386/kernel/signal.c
>
> Thanks for all,
>
> Mikael Chambon
--
Kasper Dupont
------------------------------
From: Rolf Magnus <[EMAIL PROTECTED]>
Subject: Re: usleep() is unreliable when sleeping for less then 10000 micro
Date: Mon, 9 Apr 2001 15:46:06 +0200
Pasztor Szilard wrote:
> Norm:
>> Investigate the Real-Time Linux system at
>> http://www.rtlinux.org
>>
>> You'll find the answers to all of your questions.
>
> The applications often must run on any standard linux distribution, and in
> this case, rtlinux is not a viable alternative.
man nanosleep
....
The current implementation of nanosleep is based on the
normal kernel timer mechanism, which has a resolution of
1/HZ s (i.e, 10 ms on Linux/i386 and 1 ms on Linux/Alpha).
Therefore, nanosleep pauses always for at least the speci�
fied time, however it can take up to 10 ms longer than
specified until the process becomes runnable again. For
the same reason, the value returned in case of a delivered
signal in *rem is usually rounded to the next larger mul�
tiple of 1/HZ s.
As some applications require much more precise pauses
(e.g., in order to control some time-critical hardware),
nanosleep is also capable of short high-precision pauses.
If the process is scheduled under a real-time policy like
SCHED_FIFO or SCHED_RR, then pauses of up to 2 ms will be
performed as busy waits with microsecond precision.
------------------------------
From: Markus Fischer <[EMAIL PROTECTED]>
Subject: how to modify dynamic lib ?
Date: Mon, 09 Apr 2001 15:54:06 +0200
I would like to replace a function in a dynamic lib.
assumed I have a mylibc.so.6 file and nm
shows me the functions in the library.
how can I compile and add/replace the function in
the original dynamic library ?
pointers welcome
thanks,
robert arctor wrote:
>
> Hi
>
> Could somebody please post how to permanently change the PATH environment
> variable.
>
> Thanks in advance.
>
> Rob
------------------------------
Crossposted-To: comp.os.linux.misc,comp.os.linux.setup
Subject: Suppressing Redhat bootup output
From: [EMAIL PROTECTED] (Paul Haley)
Date: Mon, 09 Apr 2001 14:33:31 GMT
Hello,
I'm running Redhat 6.2 and would like to suppress the output on bootup, such
as "Loading cron..... [OK]", etc. All I want is to have lilo say "loading
linux" and then give me a login prompt, though of course I still want all
the programs/processes to run, I just don't want them to output to the
screen. I checked out the init scripts and nothing jumped out at me.
Can somebody help?
Regards,
Paul
------------------------------
From: Mikael Chambon <[EMAIL PROTECTED]>
Subject: ALIGN in entry.S
Date: Mon, 09 Apr 2001 16:36:37 +0200
Hello,
I need a advice on a piece of assembly code in arch/i386/kernel/entry.S
====================
signal_return:
sti
testl $(VM_MASK),EFLAGS(%esp)
movl %esp,%eax
jne v86_signal_return
xorl %edx,%edx
call SYMBOL_NAME(do_signal)
jmp restore_all
ALIGN
======================
What does ALIGN mean ??
Thanks for all
--
Mikael Chambon
------------------------------
From: "Chris West" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc,comp.os.linux.setup
Subject: Re: Suppressing Redhat bootup output
Date: Mon, 9 Apr 2001 15:40:07 +0100
If you're using lilo try adding
append="quiet"
to the configuration file (usually /etc/lilo.conf) under the image you want
to be quiet.
Remember to run /sbin/lilo before rebooting.
"Paul Haley" <[EMAIL PROTECTED]> wrote in message
news:%KjA6.2341$[EMAIL PROTECTED]...
> Hello,
>
> I'm running Redhat 6.2 and would like to suppress the output on bootup,
such
> as "Loading cron..... [OK]", etc. All I want is to have lilo say
"loading
> linux" and then give me a login prompt, though of course I still want all
> the programs/processes to run, I just don't want them to output to the
> screen. I checked out the init scripts and nothing jumped out at me.
>
> Can somebody help?
>
> Regards,
> Paul
------------------------------
From: Kasper Dupont <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc,comp.os.linux.setup
Subject: Re: Suppressing Redhat bootup output
Date: Mon, 09 Apr 2001 14:48:52 +0000
This is a multi-part message in MIME format.
==============55870E7A2CA7274571662124
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Paul Haley wrote:
>
> Hello,
>
> I'm running Redhat 6.2 and would like to suppress the output on bootup, such
> as "Loading cron..... [OK]", etc. All I want is to have lilo say "loading
> linux" and then give me a login prompt, though of course I still want all
> the programs/processes to run, I just don't want them to output to the
> screen. I checked out the init scripts and nothing jumped out at me.
>
> Can somebody help?
>
> Regards,
> Paul
Why don't you want the output?
With output to the screen you can see how
far the startup has come, and in case
something goes wrong you will have a chance
to find the problem.
If you really want to do this change the
console to null in the file /etc/lilo.conf
and run /sbin/lilo. Next time you boot you
will get no output.
See my previous posting for details.
--
Kasper Dupont
==============55870E7A2CA7274571662124
Content-Type: message/rfc822
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
Message-ID: <[EMAIL PROTECTED]>
Date: Mon, 02 Apr 2001 07:19:51 +0000
From: Kasper Dupont <[EMAIL PROTECTED]>
Organization: daimi.au.dk
X-Mailer: Mozilla 3.04 (X11; I; Linux 2.2.16-3 i686)
MIME-Version: 1.0
Newsgroups: comp.os.linux.development.system
Subject: Re: Quiet Boot for Linux
References: <[EMAIL PROTECTED]>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Benjamin Scherrey wrote:
>
> I've got a techno-phobic client who gets concerned when he sees all the
> things going across his screen when linux boots up. This is enough of an
> issue that I'm prepared to go in and build a customer kernel or whatever
> is necessary to make a linux boot appear even simpler than windows. How
> do I get control over what gets to the screen when linux is booting? Is
> there some parameter I can pass while building the kernel to suppress its
> screen output? Perhaps route it to /dev/null or something?
>
> thanx & later,
>
> Ben Scherrey
All you have to do is give "console=null" as argument to
the kernel. The kernel argument can be specified in your
loader configuration. If you use lilo the configuration
is in /etc/lilo.conf. Find the default option and make a
copy, then edit the append option. When the new config
is saved run /sbin/lilo.
Example, if the default looks like this:
image=/boot/vmlinuz
label=linux
read-only
root=/dev/hda5
append="hdc=ide-scsi"
You can change it into this:
image=/boot/vmlinuz
label=linux
read-only
root=/dev/hda5
append="hdc=ide-scsi console=null"
image=/boot/vmlinuz
label=linux-old
read-only
root=/dev/hda5
append="hdc=ide-scsi"
In some distributions it can also be specified when you
install the system, but no need to reinstall just to
change this.
--
Kasper Dupont
==============55870E7A2CA7274571662124==
------------------------------
From: Kasper Dupont <[EMAIL PROTECTED]>
Subject: Re: ALIGN in entry.S
Date: Mon, 09 Apr 2001 14:56:24 +0000
Mikael Chambon wrote:
>
> Hello,
>
> I need a advice on a piece of assembly code in arch/i386/kernel/entry.S
>
> ====================
> signal_return:
> sti
> testl $(VM_MASK),EFLAGS(%esp)
> movl %esp,%eax
> jne v86_signal_return
> xorl %edx,%edx
> call SYMBOL_NAME(do_signal)
> jmp restore_all
>
> ALIGN
> ======================
>
> What does ALIGN mean ??
>
> Thanks for all
> --
> Mikael Chambon
It means that there will be added pad bytes
up to an address divisible by some number.
I don't know the number, but I think it must
be 4, 16 or 256. It is in most cases done to
improve performance.
--
Kasper Dupont
------------------------------
From: Markus Fischer <[EMAIL PROTECTED]>
Subject: how to modify dynamic lib ?
Date: Mon, 09 Apr 2001 17:06:48 +0200
I would like to replace a function in a dynamic lib.
assumed I have a mylibc.so.6 file and nm
shows me the functions in the library.
how can I compile and add/replace the function in
the original dynamic library ?
pointers welcome
thanks,
Mark
------------------------------
From: Kasper Dupont <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc
Subject: Re: process becomes [process]
Date: Mon, 09 Apr 2001 15:09:59 +0000
Chris Lo wrote:
>
> Hi,
>
> We run RH6.2 with Oracle. For most Oracle server processes, they all
> look
> alikie
>
> ora_pmon_SID
> ora_smon_SID
>
> but sometimes when the server is busy, the ps will only list them as
> [oracle].
>
> This fails our monitoring scripts. Has anyone come across this and have
> a fix for
> it?
>
> Thanks & Regards,
> Chris
At least I can tell you, that this happens when
the process is currently swapped out from memory
to disk.
I think the information that ps usually shows
about a process is part of the process memory
being swapped to disk, the system would have to
swap in a part of each process to give full
answers to the requests from ps.
It has nothing to do with the ps command it is
the proc virtual filesystem that does not report
the full info. I don't know if proc has an option
to always read the information from disk when
requested through /proc/<pid>/cmdline.
--
Kasper Dupont
------------------------------
From: Mikael Chambon <[EMAIL PROTECTED]>
Subject: Re: ALIGN in entry.S
Date: Mon, 09 Apr 2001 17:35:10 +0200
Kasper Dupont wrote:
>
> Mikael Chambon wrote:
> >
> > Hello,
> >
> > I need a advice on a piece of assembly code in arch/i386/kernel/entry.S
> >
> > ====================
> > signal_return:
> > sti
> > testl $(VM_MASK),EFLAGS(%esp)
> > movl %esp,%eax
> > jne v86_signal_return
> > xorl %edx,%edx
> > call SYMBOL_NAME(do_signal)
> > jmp restore_all
> >
> > ALIGN
> > ======================
> >
> > What does ALIGN mean ??
> >
> > Thanks for all
> > --
> > Mikael Chambon
>
> It means that there will be added pad bytes
> up to an address divisible by some number.
> I don't know the number, but I think it must
> be 4, 16 or 256. It is in most cases done to
> improve performance.
>
> --
> Kasper Dupont
Thanks Kasper
--
Mikael Chambon | Paris France | ICQ 10249913
HOME => [EMAIL PROTECTED]
WORK => [EMAIL PROTECTED] or call (651) 415 4299
SCHOOL => [EMAIL PROTECTED]
------------------------------
From: Pasztor Szilard <[EMAIL PROTECTED]>
Subject: Re: usleep() is unreliable when sleeping for less then 10000 micro
Date: 9 Apr 2001 16:13:43 GMT
Rolf:
> man nanosleep
>
> ....
> The current implementation of nanosleep is based on the
> normal kernel timer mechanism, which has a resolution of
> 1/HZ s (i.e, 10 ms on Linux/i386 and 1 ms on Linux/Alpha).
> Therefore, nanosleep pauses always for at least the speci�
> fied time, however it can take up to 10 ms longer than
> specified until the process becomes runnable again. For
> the same reason, the value returned in case of a delivered
> signal in *rem is usually rounded to the next larger mul�
> tiple of 1/HZ s.
>
> As some applications require much more precise pauses
> (e.g., in order to control some time-critical hardware),
> nanosleep is also capable of short high-precision pauses.
> If the process is scheduled under a real-time policy like
> SCHED_FIFO or SCHED_RR, then pauses of up to 2 ms will be
> performed as busy waits with microsecond precision.
No matter if it's realtime priority, it won't get the cpu back until the
scheduler has provided the deserved jiffy.
------------------------------------------------
| If you turn on the light quickly enough, |
| you can see what the dark looks like. |
------------------------------------------------
------------------------------
Crossposted-To:
24hoursupport.helpdesk,alt.comp.shareware.programmer,comp.editors,comp.lang.java.help,comp.lang.java.programmer,comp.lang.java.softwaretools,comp.os.linux.advocacy
Subject: Re: Need your recommendation for a full-featured text editor
From: Greg Copeland <[EMAIL PROTECTED]>
Date: 09 Apr 2001 11:45:18 -0500
Knowledge Seeker <[EMAIL PROTECTED]> writes:
Didn't read all of your requiremnets, but EMACS is THE editor. Like it
or not, it what all other editors try to be. Period. That doesn't mean
that it's for you, as it does have a rather steep learning curve, however,
it is THE programmer's editor.
Greg
> Hi,
>
> I am looking for a full-featured yet easy to use text editor to
> replace Notepad. Ideally, I would like a tool that is cross-platform
> or has versions for Win98 and Linux. The main platform requirement
> would be Win98 and the nice-to-have platform is Linux.
>
> There seem to be a plethora of choices that might work:
> UltraEdit
> TextPad
> EditPlus
> WinEdit
> Multi-Edit
> Zeus
> CRiSP
[snip]
--
Greg Copeland, Principal Consultant
Copeland Computer Consulting
==================================================
PGP/GPG Key at http://www.keyserver.net
DE5E 6F1D 0B51 6758 A5D7 7DFE D785 A386 BD11 4FCD
==================================================
------------------------------
From: "Dirk Riebesell" <[EMAIL PROTECTED]>
Subject: DEV TCP/IP C++
Date: Mon, 9 Apr 2001 19:23:20 +0200
Hi,
Are there any URL/HOWTO/Documentation about TCP/IP function for C++ under
Linux.
Under Win it documentet by ..DN, but for Linux I haven't found yet anything.
Thanks for any help
Dirk
------------------------------
From: [EMAIL PROTECTED] (Matthew Gatto)
Subject: Re: setitimer() problem on linux
Date: Mon, 09 Apr 2001 17:28:55 GMT
Kasper Dupont [[EMAIL PROTECTED]] wrote:
>Matthew Gatto wrote:
>>
>> I wrote a simple process scheduling shell for school, that I developed
>> on a SUN-OS Unix server. The program works fine there, but on my home
>> linux computer, it is hanging during a setitimer() call.
>>
>> After selecting a child process to give a timeslice to, it sends it
>> the start signal then does the following code (that should pause a
>> short time, and then wake up and continue executing the process
>> scheduler).
>>
>> the struct itimerval was initialized as follows at the begining of the
>> program:
>> struct itimerval new;
>> new.it_interval.tv_usec = 0;
>> new.it_interval.tv_sec = 0;
>> new.it_value.tv_usec = 600;
>> new.it_value.tv_sec = 0;
>>
>> and the sigaction initialised as follows:
>> struct sigaction action;
>> sigfillset(&(action.sa_mask));
>> action.sa_handler=do_nothing; /* do_nothing(), is as the name
>> indicates, an empty function */
>> sigaction(SIGALRM, &action, NULL);
>>
>> and during the scheduling loop the following code gets executed:
>>
>> /* start the timer */
>>
>> if (DEBUG ==1)
>> { printf("about to set itimer\n"); }
>> setitimer(ITIMER_REAL, &new, NULL);
>>
>> if (DEBUG == 1)
>> { printf("itimer set, about to pause()!"); }
>> pause();
>>
>> it hangs consistently on the first or second time (on linux at least)
>> through this code block. it will print out
>>
>> $
>> about to set itimer
>> itimer set, about to pause()!
>> about to set itimer
>>
>> and then hang indefinately.. not responding to any signals
>> (SIGINT,SIGTSTP,SIGCONT,etc). if I do a ps -el on another terminal, it
>> shows the program's WCHAN as doing a "pause". does anyone know why
>> this could be? or what i could be doing something wrong? i am truly
>> clueless.
>>
>> I posted the code here:
>> http://www.ductape.net/~gatto/scheduler/
>> The timer call that hangs is in
>> http://www.ductape.net/~gatto/scheduler/main.c
>
>You forgot to initialize the sa_flags.
>I added this line and then it worked for me:
>
> action.sa_flags=0;
>
>--
>Kasper Dupont
That was it! I was trying to figure it out for days.. Thanks very
much.
Matthew Gatto
------------------------------
** 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 by posting to the
comp.os.linux.development.system newsgroup.
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
******************************
