Linux-Development-Sys Digest #983, Volume #6 Thu, 22 Jul 99 15:13:51 EDT
Contents:
Re: basic_string on Linux using SMP - bug? (Martin von Loewis)
HEEEEELP IMLIB!!!!!!!!!!! (Guillermo)
HELP: how to measure hard disk access performance on Linux? (Dorin-Ioan MARINCA)
Re: High load average, low cpu usage (Ole Jacob Taraldset)
Re: Transparent Proxy and UDP (Andi Kleen)
Re: Tracing system calls (Gary Momarison)
Strange problem with passwords in RedHat 6.0 ("Joe")
Re: Strange problem with passwords in RedHat 6.0 (Frank v Waveren)
File size restriction on Linux : (REX)
Trying to get my optiquest v73 and my v3-3000 to work (Mike Giebel)
Current status of any Sun 3/xxx port of Linux????? (BSD Bob)
Re: File size restriction on Linux : (Greg White)
----------------------------------------------------------------------------
From: Martin von Loewis <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.development.apps
Subject: Re: basic_string on Linux using SMP - bug?
Date: 22 Jul 1999 10:47:11 +0200
"Paul Archard" <[EMAIL PROTECTED]> writes:
> Thank you for a detailed and enlightening response. My assertion
> that the standard dictates thread safety for creation is actually
> not correct - i was remembering documentation from SGI that talked
> about the thread safety in their implementation (the basis of the
> one used by egcs).
Please have a look at the egcs (and egcs-bugs) mailing list archives
at gcc.gnu.org. Several people reported that the g++ string class is
not thread-safe, although nobody could point to the exact error, or
provide a fix.
If you do find a fix (if possible, a portable one), please make sure
to submit it to [EMAIL PROTECTED]
Most other people in your situation are using the SGI string class
instead of the one provided by g++ (which does *not* come from SGI).
The problem with the SGI class is that it always copies the string on
assignment, rather than sharing it - but that might be acceptable, as
that implementation is reportedly thread-safe. You may also want to
look at the <string> implementation of libstdc++ v3.
Regards,
Martin
------------------------------
From: Guillermo <[EMAIL PROTECTED]>
Subject: HEEEEELP IMLIB!!!!!!!!!!!
Date: Thu, 22 Jul 1999 13:04:57 +0200
I am doing an image shower in gtk and using the imlib to show medical,
images in a special B&W monitor (2048x2560).
It is not a really complicated program so it has to be so quick as
possible and I have the next problems.
1)I have done a contrast in real time but the new image doesn't overlap
on the old one so my swap overflows.
2) The contrast it's no really fast (no real time working with 512x512
image) despite I work in a Sparc-20 (160 Mb Sol2.6). Besides I ONLY have
to show B&W medical images (8 bits depth) but imlib treats them as color
images, using 3 times more information. I didn't change the imrc file (I
don't know what parameters move).
Probably I have a crapy programming :-))) (look it below)
Thanks a million.
Guillermo
#ifdef HAVE_CONFIG_H
# include <config.h>
#endif
#include <gtk/gtk.h>
#include "gladesrc.h"
#include "gladesig.h"
#include <gtk/gtk.h>
#include <gdk_imlib.h>
GdkImlibImage *image;
GdkPixmap *pixmap;
GtkWidget *gtk_pixmap;
gint i=130;
GtkWidget *window;
GtkWidget *button;
GtkWidget *wc;
gint width, height;
/*-------------------------
Delete event
--------------------------*/
gint delete_event (GtkWidget *widget)
{
exit(0);
}
/*-------------------------------
Funcion del boton de contraste
--------------------------------*/
void
on_button1_clicked (GtkButton *button,
gpointer user_data)
{
GdkImlibColorModifier mod;
gtk_widget_destroy(GTK_WIDGET(gtk_pixmap));
gdk_imlib_get_image_modifier(image,&mod);
mod.gamma=mod.gamma;
mod.brightness=i;
g_print("%f\n",(double)mod.brightness);
mod.contrast=i;/*mod.contrast;*/
gdk_imlib_set_image_modifier(image,&mod);
i=i+10;
gdk_imlib_render(image,512,512);
pixmap = gdk_imlib_move_image(image);
gtk_pixmap = gtk_pixmap_new(pixmap, NULL);
gtk_container_add (GTK_CONTAINER(window), gtk_pixmap);
gtk_widget_show (gtk_pixmap);
}
/*-------------------------
contrast Button
--------------------------*/
GtkWidget*
create_wc ()
{
GtkWidget *wc;
GtkWidget *button1;
wc = gtk_window_new (GTK_WINDOW_TOPLEVEL);
gtk_widget_set_usize (wc, 100, 100);
gtk_window_set_title (GTK_WINDOW (wc), "Contraste");
gtk_window_set_policy (GTK_WINDOW (wc), FALSE, FALSE, FALSE);
button1 = gtk_button_new_with_label ("Contraste");
gtk_widget_show (button1);
gtk_container_add (GTK_CONTAINER (wc), button1);
gtk_signal_connect_object (GTK_OBJECT (button1), "clicked",
GTK_SIGNAL_FUNC (on_button1_clicked),
GTK_OBJECT (wc));
return wc;
}
/*-------------------------
Main
--------------------------*/
int
main (int argc, char *argv[])
{
gtk_set_locale ();
gtk_init (&argc, &argv);
/* Imlib initialization fxns */
gdk_imlib_init();
gtk_widget_push_visual(gdk_imlib_get_visual());
gtk_widget_push_colormap(gdk_imlib_get_colormap());
image = gdk_imlib_load_image("PPM");
width = image->rgb_width;
height = image->rgb_height;
gdk_imlib_render(image,512,512);
pixmap = gdk_imlib_move_image(image);
// event_box = gtk_event_box_new();
window = gtk_window_new(GTK_WINDOW_TOPLEVEL);
/* I'm just ignoring the Mask for now. */
gtk_pixmap = gtk_pixmap_new(pixmap, NULL);
gtk_window_set_title (GTK_WINDOW (window), "Isto fixeno eu");
gtk_signal_connect (GTK_OBJECT (window), "delete_event",
GTK_SIGNAL_FUNC(gtk_widget_destroy),
GTK_OBJECT(window));
gtk_signal_connect (GTK_OBJECT (window), "destroy",
GTK_SIGNAL_FUNC(gtk_widget_destroy),
GTK_OBJECT(window));
gtk_window_set_policy (GTK_WINDOW (window), FALSE, FALSE, FALSE);
gtk_container_add (GTK_CONTAINER(window), gtk_pixmap);
gtk_widget_show (gtk_pixmap);
/* Now show our image */
gtk_widget_show(window);
wc = create_wc ();
gtk_widget_show (wc);
gtk_main ();
return 0;
}
------------------------------
From: Dorin-Ioan MARINCA <[EMAIL PROTECTED]>
Subject: HELP: how to measure hard disk access performance on Linux?
Date: Thu, 22 Jul 1999 15:38:30 +0200
Hi!
I want to know if exists some utils that can be used to obtain
performances for hard disks on Linux. I'm interested also for networking
subsystem.
Or maybe better... where can I obtain an updated doc for 'proc'
filesystem infos? Because there are some fields which maybe can be used
to calculate disk throughput [Mbps]:
[marinca@devl67 /proc]$ cat stat
cpu 145829 42 54336 10545351
disk 40282 85451 40232 0 #?
disk_rio 40282 50086 40232 0 #?
disk_wio 0 35365 0 0 #? what are these?
disk_rblk 40284 111122 40234 0 #?
disk_wblk 0 84860 0 0 #?
page 255216 90741
swap 1825 4131
intr 11778803 10745558 89154 0 67832 0 16023 20 0 1 0 0 0 695057 1
124922 40235
ctxt 6936714
btime 932542995
processes 5798
I'm interested on IDE & SCSI disks. And for network: Ethernet & ATM...
I have a standard RH52 Linux.
Thanks for any little hint,
D.I.M.
------------------------------
From: Ole Jacob Taraldset <[EMAIL PROTECTED]>
Subject: Re: High load average, low cpu usage
Date: Thu, 22 Jul 1999 15:16:04 +0200
I think maybe it has something to do with the /home partition being NFS
mounted. Since NFS is implemented in kernel space in 2.2 is it possible
that NFS is struggling and it doesn't show up as a process in top/ps? Is
there some way to figure this out? I can't find anything except lines like
these in "/var/log/messages":
Jul 22 02:57:27 ojt kernel: __nfs_fhget: inode 0 busy, i_count=2,
i_nlink=1
Jul 22 02:57:27 ojt kernel: nfs_free_dentries: found //idar, d_count=0,
hashed=1
Jul 22 02:57:27 ojt kernel: nfs_dentry_delete: //idar: ino=0, count=2,
nlink=1
Any ideas?
Ole Jacob
On Thu, 22 Jul 1999, Joseph Mack wrote:
> Ole Jacob Taraldset wrote:
> >
> > I have a PII 450 MHz w/512 MB ram and a clean RedHat 6.0 w/updates .
> > When I take a look at cpu usage i kpm/qps most of the cpu is idle
> > (~85%), but load average reports around 2. Isn't load average a function
> > of cpu usage (only, mostly)?
>
> This does sound a little unusual, however the load average is the number
> of requests queued in the kernel. If the requests are to slow devices
> (disks and modem lines are the worst) and the CPU is waiting for them
> then you'll get high load averages and low CPU use
>
> Joe
> --
> Joseph Mack PhD, Senior Systems Engineer, Lockheed Martin
> contractor to the National Environmental Supercomputer Center,
> mailto:[EMAIL PROTECTED] ph# 919-541-0007, RTP, NC, USA
>
GexCon AS, Bergen, Norway <http://www.gexcon.com>
Tel : +47 55574334 (office) +47 55558650 (home)
Mob.tel. : +47 95080525 Fax : +47 55574331
PGP key : <http://home.c2i.net/ojt/pgp.txt>
------------------------------
From: Andi Kleen <[EMAIL PROTECTED]>
Subject: Re: Transparent Proxy and UDP
Date: 22 Jul 1999 16:41:43 +0200
Peter Eisenlohr <[EMAIL PROTECTED]> writes:
> Hi all,
>
> I am redirecting some UDP traffic via an appropiate ipchains-rule to
> some local port. With recvfrom(), I can receive datagrams and identify
> their source(s). But how can I determine the destination, i.e. the
> IP-Address the packet has been sent to? (ipchains will catch the
> packet regardless of it's destination address and send it to my
> program)
transparent uses a superugly hack to archieve this and puts the
destination address in 8 unused bytes in sockaddr_in at the end.
You have to set the MSG_PROXY flag.
So use something like that:
struct sockaddr_in adr, *to;
int alen=sizeof adr;
recvfrom(...., MSG_PROXY,&adr,&alen);
to = (struct sockaddr_in *) &adr.sin_zero;
For sending it works similar, just that sin_zero can forge the
from address if MSG_PROXY is set.
Yes, it smells. Linux 2.4 will have a proper solution.
-Andi
---
This is like TV. I don't like TV.
------------------------------
Crossposted-To: comp.os.linux.development,comp.os.linux.misc
Subject: Re: Tracing system calls
From: Gary Momarison <[EMAIL PROTECTED]>
Date: 22 Jul 1999 09:35:18 -0700
Mitch <[EMAIL PROTECTED]> writes:
> I'm currently working on a BSDi box, and have recently been introduced
> to a program called 'ktrace'. This thing is a great utility. My
> question is whether or not there is something like this for linux? I've
> made a cursory pass through freshmeat, but didn't turn up anything.
http://www.aa.net/~swear/pedia/debugging.html had this link:
http://www.linuxgazette.com/issue01to08/lg_issue6.html#strace
to an article about "strace" which is probably similar to ktrace.
--
Look for Linux info at http://www.dejanews.com/home_ps.shtml and
Gary's Encyclopedia at http://www.aa.net/~swear/pedia/index.html
------------------------------
From: "Joe" <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.linux.misc,linux.redhat.devel,linux.redhat.install,linux.redhat.pam,linux.redhat.sparc
Subject: Strange problem with passwords in RedHat 6.0
Date: Thu, 22 Jul 1999 15:57:05 GMT
Hi,
I'm using RedHat 6.0 Kernel 2.2.5-15 on a SPARCserver 5. The installation
was downloaded from gd.tuwien.acat in vienna. After compiling the kernel for
IP masquerading, I've tried to install a mailserver. Sendmail worked fine
for the first, but was not very comfortable, when I was out of office. So I
decided to install the qpopper POP3 mailserver, because it seems to be one
of the best (the best?). Installation was very simple, and my first tests
(telent to server over pop3 port) were successful. But when a user tries to
connect, he always get he message wrong password. So I tried different
configurations (spcial-auth parametor for configure, -DAUTH in makefile) and
a lot of the stuff I found in FAQ's and the suggestions people (THANKS!!)
sent me via e-mail.
Getting frustrated by beeing unable to get it work in 10 days, I've tried to
use Cyrus IMAP server instead. Download, configuration, compile and install
were also very simple, the first test (telent.... see above) and the test
utility imtest shows me, that I'm on the right way. But when I tried to use
cyradm the program asked me for a user (I tried root) and a password (as you
guessed, tried the root-password). Got the message Wrong Password !! (...
pwcheck is running, tried several configuration options, bla,bla bla....)
It seems to me (correct me, if I'm wrong) that the pwcheck program compares
the password which has been entered in the mail-client with the one who is
in the /etc/shadow file ( I AM using shadow password file !!).
After adding some debug code (very simple printf commands) in
pwcheck_getspnam.c (also tried the pwcheck_getpwnam.c version, but the
passwords which will be compared are that different, so I think this is the
wrong way...) I got the following output:
getspnam-Debug:
Password from /etc/shadow : $1$T1gp7QPv$QeJ278puZM3B.zDIUA.Nk1
Password generated by crypt() : $1$T1gp7QPv$3gmKod9BErPhayuKWvmLV1
The crypted password from /etc/shadow is O.K. But the password which will be
generated from the plaintext password is different from the one in the
shadow file.
BUT WHY. Don't they use the same mechanism to decrypt passwords ?????
I would be very thankful for EVERY help I get. I'm stuck now, and do not
know what else to try....
a frustrated
Joe
------------------------------
From: [EMAIL PROTECTED] (Frank v Waveren)
Subject: Re: Strange problem with passwords in RedHat 6.0
Crossposted-To:
comp.os.linux.misc,linux.redhat.devel,linux.redhat.install,linux.redhat.pam,linux.redhat.sparc
Date: Thu, 22 Jul 1999 17:21:05 GMT
In article <lbHl3.18028$[EMAIL PROTECTED]>,
"Joe" <[EMAIL PROTECTED]> writes:
> Hi,
>
> I'm using RedHat 6.0 Kernel 2.2.5-15 on a SPARCserver 5. The installation
> was downloaded from gd.tuwien.acat in vienna. After compiling the kernel for
> IP masquerading, I've tried to install a mailserver. Sendmail worked fine
> for the first, but was not very comfortable, when I was out of office. So I
> decided to install the qpopper POP3 mailserver, because it seems to be one
> of the best (the best?). Installation was very simple, and my first tests
> (telent to server over pop3 port) were successful. But when a user tries to
> connect, he always get he message wrong password. So I tried different
> configurations (spcial-auth parametor for configure, -DAUTH in makefile) and
> a lot of the stuff I found in FAQ's and the suggestions people (THANKS!!)
> sent me via e-mail.
> Getting frustrated by beeing unable to get it work in 10 days, I've tried to
> use Cyrus IMAP server instead. Download, configuration, compile and install
> were also very simple, the first test (telent.... see above) and the test
> utility imtest shows me, that I'm on the right way. But when I tried to use
> cyradm the program asked me for a user (I tried root) and a password (as you
> guessed, tried the root-password). Got the message Wrong Password !! (...
> pwcheck is running, tried several configuration options, bla,bla bla....)
>
> It seems to me (correct me, if I'm wrong) that the pwcheck program compares
> the password which has been entered in the mail-client with the one who is
> in the /etc/shadow file ( I AM using shadow password file !!).
>
> After adding some debug code (very simple printf commands) in
> pwcheck_getspnam.c (also tried the pwcheck_getpwnam.c version, but the
> passwords which will be compared are that different, so I think this is the
> wrong way...) I got the following output:
>
> getspnam-Debug:
> Password from /etc/shadow : $1$T1gp7QPv$QeJ278puZM3B.zDIUA.Nk1
> Password generated by crypt() : $1$T1gp7QPv$3gmKod9BErPhayuKWvmLV1
>
> The crypted password from /etc/shadow is O.K. But the password which will be
> generated from the plaintext password is different from the one in the
> shadow file.
>
> BUT WHY. Don't they use the same mechanism to decrypt passwords ?????
>
> I would be very thankful for EVERY help I get. I'm stuck now, and do not
> know what else to try....
Just ignoring you're little debug-quest, have you tried a non-root user? lotsa
software won't allow plaintext sending of the root pw (or atleast not accept
it).
--
Frank v Waveren
[EMAIL PROTECTED]
ICQ# 10074100
------------------------------
From: REX <[EMAIL PROTECTED]>
Subject: File size restriction on Linux :
Date: Thu, 22 Jul 1999 14:15:55 GMT
Hi,
I am using linux kernel 2.2.5 (SuSE 6.1)
I want to create a file of size 4 Gbytes. I tried to do it with the
following command :
dd if=/dev/zero of=/home/xxx.txt count=4000000 ibs=1024
The process stops after creating the file of size 2147482648 which is
nothing but 2 Gbytes.
Though linux howto documents does not say any restriction on the files
regarding the file size (I am also refering to a book on Linux Kernel), I
don't know why am I not alloed to create such a big file.
Is there any way of acieving it or that's a limitation??
Thanx in advance.
Regards,
Rekha Kulkarni.
------------------------------
From: Mike Giebel <[EMAIL PROTECTED]>
Subject: Trying to get my optiquest v73 and my v3-3000 to work
Date: Thu, 22 Jul 1999 11:52:46 -0500
I'm trying to get Gnome to work on my Optiquest v73 and my v3-3000....
so far i am able to get gnome to work BUT i can only see a fourth of the
desktop on my
montitor in vga mode.
please help! :)
email me at [EMAIL PROTECTED]
-mav
------------------------------
From: BSD Bob <[EMAIL PROTECTED]>
Subject: Current status of any Sun 3/xxx port of Linux?????
Date: 22 Jul 1999 17:10:28 GMT
I have several old Sun 3/xxx machines in the basement, and was wondering
if any progress had been made on a Sun 3 port of Linux? Any insights
or pointers are appreciated.
Thanks
Bob Keys
------------------------------
From: Greg White <[EMAIL PROTECTED]>
Subject: Re: File size restriction on Linux :
Date: Thu, 22 Jul 1999 18:15:47 GMT
Frank v Waveren wrote:
>
> In article <[EMAIL PROTECTED]>,
> REX <[EMAIL PROTECTED]> writes:
> >
> > Hi,
> >
> > I am using linux kernel 2.2.5 (SuSE 6.1)
> >
> > I want to create a file of size 4 Gbytes. I tried to do it with the
> > following command :
> >
> > dd if=/dev/zero of=/home/xxx.txt count=4000000 ibs=1024
> >
> > The process stops after creating the file of size 2147482648 which is
> > nothing but 2 Gbytes.
> >
> > Though linux howto documents does not say any restriction on the files
> > regarding the file size (I am also refering to a book on Linux Kernel), I
> > don't know why am I not alloed to create such a big file.
> >
> > Is there any way of acieving it or that's a limitation??
>
> IIRC, linux has no problem with files > 2GB, but ext2 has. (or was it the
> other way round? Dunno, doesn't matter). I believe it's been fixed in 2.2.10-ac11
>
> --
>
> Frank v Waveren
> [EMAIL PROTECTED]
> ICQ# 10074100
It is the 32-bit kernel, not ext2. Without hacks, that's the file size
limit for standard IA-32 architecture.
GW
------------------------------
** 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
******************************
