Re: gdb bt display problem.
On Tue, Aug 20, 2002 at 05:43:23PM +0300, Shlomi Fish wrote: Hi! I'm trying to debug a gtk+ 1.2.10 application (GImageView to be specific). The problem is that when I type Ctrl+C and then type bt, I get the following display: #0 0x4049bc74 in poll () from /lib/libc.so.6 #1 0x402d9734 in g_source_remove_by_funcs_user_data () from /usr/lib/libglib-1.2.so.0 Cannot access memory at address 0x2 What can I do to fix it? Make sure everything is compiled with debug symbols and frame pointers, especially glib? It looks like gdb gets confused because the stack layout is different than what it expects. If you compiled the application from source (I assume you did, not much point trying to debug it otherwise), the path of least resistance is to debug only the application code and treat its library calls as black boxes. That means figuring out what appliation code calls the functions above (or what gets called by them) and setting a break point there explicitly. I use gdb 5.2.1 on a Mandrake Linux 8.2. (it happened with gdb 5.1.x, too) I doubt it has anything to do with gdb versions. -- calm down, it's *only* ones and zeros. http://syscalltrack.sf.net/ http://vipe.technion.ac.il/~mulix/ msg21265/pgp0.pgp Description: PGP signature
Re: Slightly OT: yellow pages blocks our browsers
On Thu, Apr 04, 2002 at 03:16:32PM +0300, Yotam Rubin wrote: P.S.: Fellas, RTFH (Headers), do not CC me on list replies. It clogs up my spool and serves no useful purpose. man procmailex | col -b | vi - /duplicates -- The ill-formed Orange Fails to satisfy the eye: http://vipe.technion.ac.il/~mulix/ Segmentation fault. http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: c question
On Thu, Apr 04, 2002 at 06:00:36PM +0300, Shlomi Fish wrote: write(1, World!\n, 7); please use fileno(stdout). daemon writers everywhere will thank you. -- The ill-formed Orange Fails to satisfy the eye: http://vipe.technion.ac.il/~mulix/ Segmentation fault. http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: c question
On Thu, Apr 04, 2002 at 06:17:16PM +0300, Nadav Har'El wrote:
On Thu, Apr 04, 2002, mulix wrote about Re: c question:
On Thu, Apr 04, 2002 at 06:00:36PM +0300, Shlomi Fish wrote:
write(1, World!\n, 7);
please use fileno(stdout). daemon writers everywhere will thank you.
You're right about the existance of the fileno() function, but I'm not sure
I understand your advice. The only reason fileno(stdout) is *ever* going to
be different than 1, is if freopen() is used. freopen() is an even more
obscure function than the fdopen() function I mentioned in a previous
post, and I don't think I ever saw it actually used. Why is it used by
daemon writers everywhere?
I guess I'm missing some important point.
consider this snippet, in a library function:
int do_something()
{
char msg[] = an incredibly important event occured;
write(fileno(stdin), msg, sizeof(msg));
return 0;
}
and now consider this typical daemon code:
int daemonize()
{
...
/* close all open file descriptors */
...
/* open something application specific, which happens to get the fd 1 */
...
/* inditectly direct library output to our logging files */
stdout = fopen(/my/logging/file);
stderr = fopen(/my/other/logging/file);
}
now, wherever the library writer used stdout/stderr by name, whether
using the glibc IO facilities (printf(), fprintf() or by system calls,
everything will just work - unless the writer made the undue
assumption that stdout == 1. if the value 1 (or 2) is hard coded, much
fun will ensue.
as for the fact that libraries shouldn't write to stdout and stderr,
that's a completely different discussion.
--
The ill-formed Orange
Fails to satisfy the eye: http://vipe.technion.ac.il/~mulix/
Segmentation fault. http://syscalltrack.sf.net/
=
To unsubscribe, send mail to [EMAIL PROTECTED] with
the word unsubscribe in the message body, e.g., run the command
echo unsubscribe | mail [EMAIL PROTECTED]
Re: c question
On Thu, Apr 04, 2002 at 06:47:13PM +0300, Nadav Har'El wrote: On Thu, Apr 04, 2002, mulix wrote about Re: c question: On Thu, Apr 04, 2002 at 06:17:16PM +0300, Nadav Har'El wrote: stdout = fopen(/my/logging/file); stderr = fopen(/my/other/logging/file); Ok, I get your point. I don't remember seeing any program doing something like that, but I guess it's conceivable, and that using fileno() might be a wise idea. i'm pretty sure i've picked it up somewhere, perhaphs at steven's APUE. wherever it was, it's a good practice, and you get the added benefit of readability of the code. One last point though: it is not portable to assign to stdout, etc. like you showed above. In Linux you can do this, because stdout is defined as extern FILE *stdout;/* Standard output stream. */ but in other UNIX systems you can't. For example, in Solaris (8, at least), and in all ATT-based systems I knew, stdout is defined as #define stdout (__iob[1]) And you obviously cannot assign to this. You may, perhaps, do *stdout=*fopen(...); but that is really funky and again non-portable; Instead, the portably ANSI- C solution is to use the freopen() function. true. my example was scratch code - please treat it as such. -- The ill-formed Orange Fails to satisfy the eye: http://vipe.technion.ac.il/~mulix/ Segmentation fault. http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: c question
On Thu, Apr 04, 2002 at 06:03:40PM +0200, Shachar Shemesh wrote: Just wanted to quote an amuzing experience I had in the single haifux club meeting I attended. I made a comment about something only working on Linux, and having to use X in order to make it Unix compatible, and got bemused looks from everyone around me saying This is a Linux meeting. Sigh. well, speaking as a proud haifux member, some of us regularly use other operating systems, whether of the unixish kind or even (gasp!) of the windows kind. nonetheless, occasionally we discuss topics at our meeting which are linux specific - we *are* a linux club, after all. we try to label them as such, because world domination has not been achieved yet. Some people are too narrowminded. all too true, however, i doubt you'll find many haifux members on that particular list of people. looking at the list of last ten lectures, i would classify 5 as linux specific, although in a broad sense of the word, and 5 as completely linux neutral. of the next two lectures scheduled, neither is linux specific. i might as well take this opportunity to invite you all. guy keren will talk about pthreads this upcoming monday (08/04) and orna agmon will talk about PVM (parallel virtual machine) on monday, the 22/04. more details are available at http://linuxclub.il.eu.org. see you all there! -- The ill-formed Orange Fails to satisfy the eye: http://vipe.technion.ac.il/~mulix/ Segmentation fault. http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: C++: a constructor can not operate on a private static map object?
On Mon, Apr 01, 2002 at 08:09:44PM +0300, Shaul Karl wrote: [what does c++ have to do with linux-il? this question would've probably been better suited to hackers-il, or cpp-are-us-il, or bored-students-dont-wanna-study-answer-questions-all-day-in-email-il]. In what follows I fail to operate in the class constructor on a map object that was defined as a private static member of that class. Can any one points to what is being done wrong? yes. to quote the linker: /tmp/ccN1chDT.o: In function `category::category[not-in-charge](std::bas ic_stringchar, std::char_traitschar, std::allocatorchar const)': /tmp/ccN1chDT.o(.text+0x29): undefined reference to `category::allCategories' that is, you're referring to an object which doesn't have any storage allocated for it. adding this line mapconst string, category* category::allCategories; to your main.cc would solve your problem. -- The ill-formed Orange Fails to satisfy the eye: http://vipe.technion.ac.il/~mulix/ Segmentation fault. http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: ftp recursive chmod
On Fri, Mar 29, 2002 at 01:13:40PM +0200, Tzahi Fadida wrote: since no one answered i wish to rephrase and add. Can someone point me or write a simple perl script (and i know some of u can do i in 1 line :) that changes all the files attributes (ex: chmod 766) in an ftp site accross subdirectries? RTFM find (1). find ./ -name * -type f | xargs chmod 766 you can also do the same thing using find's -exec option, but i always found its syntax non intuitive. [and trim your lines at 72 characters, please]]] -- The ill-formed Orange Fails to satisfy the eye: http://vipe.technion.ac.il/~mulix/ Segmentation fault. http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: ftp recursive chmod
On Fri, Mar 29, 2002 at 01:48:48PM +0300, Guy Cohen wrote: He wants to do it over ftp. oh, missed that. well, that makes it a bit more interesting. easiest way would probably be to write a script in one's favorite scripting language which has an ftp module (perl and python do) which walks the directory structure and calls 'chmod' on each file. should be rather easy to implement (from scratch or modifying an example). -- The ill-formed Orange Fails to satisfy the eye: http://vipe.technion.ac.il/~mulix/ Segmentation fault. http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: ftp recursive chmod
On Fri, Mar 29, 2002 at 01:52:10PM +0300, mulix wrote: On Fri, Mar 29, 2002 at 01:48:48PM +0300, Guy Cohen wrote: He wants to do it over ftp. oh, missed that. well, that makes it a bit more interesting. easiest way would probably be to write a script in one's favorite scripting language which has an ftp module (perl and python do) which walks the directory structure and calls 'chmod' on each file. should be rather easy to implement (from scratch or modifying an example). [replying to myself, shame on me] #!/usr/bin/env python # walk an ftp server and apply a command to each file # # (c) [EMAIL PROTECTED], GPL. import ftplib, string eInvalidServer = 'No server specified' class FtpWalker: ftp = None visitme = [] cwd = None command = None def __init__(self, server, command = None, startdir = '/'): if (server == None): raise eInvalidServer self.command = command self.cwd = startdir self.ftp = ftplib.FTP(server) self.ftp.login() self.ftp.cwd(self.cwd) def ftp_callback(self, str): print (callback got: %s % str) details = string.split(str) if (str[0] == 'd'): # directory, schedule a visit name = details[8] print got a directory: %s % name self.visitme.append(self.cwd + name) elif (str[0] == '-'): # regular file, apply the command to it name = details[8] print got a file: %s % name if (self.command != None): cmd = (self.command % name) print sending: %s % cmd resp = self.ftp.sendcmd(cmd) print %s returned %s % (cmd, resp) def walk(self): self.ftp.dir(self.ftp_callback) while (len(self.visitme)): try: newdir = self.visitme.pop() self.ftp.cwd(newdir) self.cwd = newdir + '/' self.ftp.dir(self.ftp_callback) except ftplib.error_perm, e: print e # but continue if __name__ == __main__: server = 'ftp.iglu.org.il' # name of your server here command = 'SIZE %s' # (or whatever you want - %s for the file name) startdir = '/' # of /path/to/my/directory f = FtpWalker(server, command, startdir) f.walk() -- The ill-formed Orange Fails to satisfy the eye: http://vipe.technion.ac.il/~mulix/ Segmentation fault. http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: ulimit -c
On Thu, Mar 21, 2002 at 02:54:06PM +0200, Christoph Bugel wrote: For some reason, by default, my ulimit -c is 0, It means no corefiles are generated. I think I'll add ulimit -c unlimited to my .bash_login Or would it be a bad idea? only if you run huge processes which might leave tens or hunderds of megabytes of corefiles lying around, and diskspace is scarce. -- The ill-formed Orange Fails to satisfy the eye: http://vipe.technion.ac.il/~mulix/ Segmentation fault. http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: tal@whatsup.co.il does not work
On Tue, Mar 19, 2002 at 07:30:45PM +0200, Amir Tal wrote: It does work..i just sent myself several tests from yahoo, hotmail and my account at work. All good. Can you be more specific about the error you are getting ? [explanation sent in private. it has to do with the fact that whatsup.co.il used to be hosted at actcom and actcom wasn't notified that it's no longer hosted there]. -- The ill-formed Orange Fails to satisfy the eye: http://vipe.technion.ac.il/~mulix/ Segmentation fault. http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: pthreads question
On Mon, Mar 18, 2002 at 01:52:36PM +0200, Malcolm Kavalsky wrote:
Eureka!
Nadav Har'El wrote:
On Mon, Mar 18, 2002, Malcolm Kavalsky wrote about Re: pthreads
question:
I asked one of the top Unix hackers that I know, and he said:
I would guess that if you do large af_unix transfers that are page
aligned then the system doesn't have to actually copy the data rather it
can share the page and do a copy on write. This preserves the socket
semantics and can be faster than memcpy. This was done many years ago in
Solaris.
I wonder if digging deep enough in the kernel sources, will reveal
this ...
You can try to check if this is the case, by following each send or
memcpy
by a memset() of the buffer. If the memcpy method suddenly becomes
quicker,
this explanation might be true.
Strange though - how come malloc() returns page-aligned buffers? Does the
Linux code really checks for this rare and rather esoteric case (if you
write to the buffer after sending it, and the kernel can't know you're
writing whole pages, it will have to do a copy-on- write and do the copy
anyway).
This is exactly what happened! I added in memset after memcpy, and also
after sending the buffer, the results are:
Memcpy'ed and memsetted 1000 blocks of size 1048576 in 18 seconds = 55
Mbytes/second
Started receiving at Mon Mar 18 13:41:13 2002
Received 1048576000 bytes in 17 seconds over unix socket = 59
Mbytes/second
Started sending at Mon Mar 18 13:41:13 2002
Sent and memsetted 1000 blocks of size 1048576 in 17 seconds over unix
socket = 58 Mbytes/second
i decided to play too. i took your code and modified it, so that the
tests are run seperately (since i didnt want the after effects from
fork's COW behaviour to affect the memcpy case). i also modified it to
use getrusage().
here are my results:
[mulix@alhambra tmp]$ for arg in 1 2 3; do ./b memcpy ; done ;
memcpy'ed 1000 blocks of size 1048576. user time: 16.07 secs, system time:
0.06 secs
memcpy'ed 1000 blocks of size 1048576. user time: 15.96 secs, system time:
0.04 secs
memcpy'ed 1000 blocks of size 1048576. user time: 15.92 secs, system time:
0.06 secs
[mulix@alhambra tmp]$ for arg in 1 2 3; do ./b send ; done ;
sent1000 blocks of size 1048576. user time: 6.99 secs, system time:
10.02 secs
sent1000 blocks of size 1048576. user time: 7.42 secs, system time:
10.33 secs
sent1000 blocks of size 1048576. user time: 7.11 secs, system time:
10.38 secs
kernel is 2.4.18rc1, and here's the modified code:
#include stdio.h
#include malloc.h
#include string.h
#include time.h
#include sys/socket.h
#include sys/time.h
#include sys/resource.h
#include sys/un.h
#include sys/types.h
#include sys/wait.h
#include unistd.h
#include assert.h
#define BUFSIZE 0x10 /* 1 Megabyte */
#define NBLOCKS 1000
#define PORT_NAME/tmp/foo
void server(), client();
void socket_benchmark()
{
pid_t rc;
if ( (rc = fork()) == 0 ) {
server();
waitpid(rc, NULL, 0);
} else {
sleep(1); /* Dirty, but ensures client runs after server is ready */
client();
}
}
void server()
{
struct sockaddr_un sin,from;
int s,g,len;
char *buf;
buf = malloc( BUFSIZE );
/* Create an unbound socket */
if( (s=socket( PF_UNIX, SOCK_STREAM, 0 )) 0 ){
perror(Bad socket\n);
return;
}
strcpy( sin.sun_path, PORT_NAME );
sin.sun_family = PF_UNIX;
if( bind( s, (struct sockaddr *)sin,
sizeof(sin) ) 0){
perror(bind);
return;
}
listen( s, 5 );
len = sizeof(from);
g = accept( s, (struct sockaddr *)from, len );
while( read( g, buf, BUFSIZE ) 0 ); /* sink all data received */
close(g);
close(s);
unlink( PORT_NAME );
}
void client()
{
struct rusage r = {{0},};
struct sockaddr_un sin;
int s;
char *buf;
time_t start_time, elapsed_time;
int i;
assert(!(r.ru_utime.tv_sec | r.ru_utime.tv_usec |
r.ru_stime.tv_sec | r.ru_stime.tv_usec));
buf = malloc( BUFSIZE );
if( (s=socket( PF_UNIX, SOCK_STREAM, 0 )) 0 ){
perror(socket);
return;
}
strcpy( sin.sun_path, PORT_NAME );
sin.sun_family = PF_UNIX;
if( connect( s, (struct sockaddr *)sin, sizeof(sin)) 0 ){
perror(connect);
close(s);
return;
}
start_time = time(0);
for( i=0; i NBLOCKS write(s, buf, BUFSIZE) == BUFSIZE ; i++ )
memset( buf, 'A', BUFSIZE );;
elapsed_time = time(0) - start_time;
close(s);
#if 0
printf
Re: pthreads question
On Mon, Mar 18, 2002 at 02:39:09PM +0200, Malcolm Kavalsky wrote: mulix wrote: i decided to play too. i took your code and modified it, so that the tests are run seperately (since i didnt want the after effects from fork's COW behaviour to affect the memcpy case). i also modified it to use getrusage(). here are my results: [mulix@alhambra tmp]$ for arg in 1 2 3; do ./b memcpy ; done ; memcpy'ed1000 blocks of size 1048576. user time: 16.07 secs, system time: 0.06 secs memcpy'ed1000 blocks of size 1048576. user time: 15.96 secs, system time: 0.04 secs memcpy'ed1000 blocks of size 1048576. user time: 15.92 secs, system time: 0.06 secs [mulix@alhambra tmp]$ for arg in 1 2 3; do ./b send ; done ; sent 1000 blocks of size 1048576. user time: 6.99 secs, system time: 10.02 secs sent 1000 blocks of size 1048576. user time: 7.42 secs, system time: 10.33 secs sent 1000 blocks of size 1048576. user time: 7.11 secs, system time: 10.38 secs Interesting, I compiled and ran your code with results: memcpy'ed 1000 blocks of size 1048576. user time: 17.83 secs, system time: 0.01 secs sent1000 blocks of size 1048576. user time: 8.19 secs, system time: 5.67 secs The sum of user and system time is pretty much equal (just as in yours) I then commented out the memset commands and got: memcpy'ed 1000 blocks of size 1048576. user time: 8.90 secs, system time: 0.04 sec sent1000 blocks of size 1048576. user time: 0.00 secs, system time: 0.62 secs This is a dramatic difference. Did you try this ? based on nadav's suggestion, i added getrusage() in the server as well. here are the results: [mulix@alhambra tmp]$ echo without memsets: ; ./b memcpy; ./b send without memsets: memcpy'ed 1000 blocks of size 1048576. user time: 7.98 secs, system time: 0.01 secs client sent 1000 blocks of size 1048576. user time: 0.01 secs, system time: 5.07 secs server received 1048576000 bytes. user time: 0.02 secs, system time: 6.75 secs [mulix@alhambra tmp]$ echo with memsets: ; ./b memcpy; ./b send with memsets: memcpy'ed 1000 blocks of size 1048576. user time: 15.54 secs, system time: 0.09 secs client sent 1000 blocks of size 1048576. user time: 7.40 secs, system time: 9.56 secs server received 1048576000 bytes. user time: 0.04 secs, system time: 7.67 secs personally, i'm very curious why the memcpy case takes so much user time. objdump to the rescue. -- The ill-formed Orange Fails to satisfy the eye: http://vipe.technion.ac.il/~mulix/ Segmentation fault. http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: [Humour] This is hilarious... ;-)
On Mon, Mar 11, 2002 at 05:49:27PM +0200, Tzafrir Cohen wrote: On Mon, 11 Mar 2002, Gilad Ben-Yossef wrote: http://www.microsoft.comitem=linux@3573468885/original.html I hate to spoil it for others, but... Is this a valid URL? it's a valid url, just not the url most people think it is. user: www.microsoft.com@item=linux website: 3573465 which translates to 212.254.206.213 -- The ill-formed Orange Fails to satisfy the eye: http://vipe.technion.ac.il/~mulix/ Segmentation fault. http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Fwd: [pptp-devel] pptp-linux-1.1.0-rc3 available for testing
a new version of pptp will be out Real Soon Now(tm). if anyone wants to test the quirks handling to verify that i works with israeli adsl modems, please do so. (it works for me). remember to run it with /pptp 10.0.0.138 --quirks=BEZEQ_ADSL pppd options here! - Forwarded message from James Cameron [EMAIL PROTECTED] - Date: Thu, 07 Mar 2002 11:05:48 +1100 From: James Cameron [EMAIL PROTECTED] Organization: Netrek Vanilla Server Maintainer X-Mailer: Mozilla 4.7 [en] (X11; I; OSF1 V4.0 alpha) To: [EMAIL PROTECTED], [EMAIL PROTECTED] Reply-To: [EMAIL PROTECTED] Subject: [pptp-devel] pptp-linux-1.1.0-rc3 available for testing G'day, Release candidate 3 of PPTP Client 1.1.0 is available for testing by the development team. Packages have been built for Debian and Red Hat on Alpha and Intel. http://quozl.netrek.org/pptp/ http://quozl.linux.org.au/pptp/ Changelog - allow activation as a psuedo-tty child process from pppd. - ADSL modem quirks handler by [EMAIL PROTECTED] - enhance bad FCS error message. - ported to FreeBSD and NetBSD. The activation as a psuedy-tty child process from pppd works for me, provided I include options 'logfd 2 nodetach', otherwise what happens is that pppd (2.4.0) diagnostic messages are sent through the pty to the GRE encapsulator. This causes the 'no GRE from server' symptom. Indicator 1: if you have considerably more GRE packets sent by the client than there are LCP packets logged by the client. Indicator 2: examining GRE packets using tcpdump -X shows they contain text messages that pppd would normally emit. -- James Cameron ___ pptpclient-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/pptpclient-devel - End forwarded message - -- The ill-formed Orange Fails to satisfy the eye: http://vipe.technion.ac.il/~mulix/ Segmentation fault. http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: mtu problems?
On Mon, Mar 04, 2002 at 08:23:06PM +0200, Aviram Jenik wrote: A question to the MTU gurus (Muli/Dani?): dani is the real expert, i'm just using a few handy heuristics. I'm pretty sure I have an MTU problem. However, I can't figure out: A. How to 'debug' it (i.e. I don't know if the problem is really MTU) pakcet dump (ethereal, tcpdump). you will see the client sending a packet and getting one packet in a response, or none at all. then it will continue sending and get no response. B. What the problem is (if it exists). see dani's note, at http://damyen.technion.ac.il/~dani/adsl-mtu.txt I think that (A) is especially important, since I'm getting the feeling I'm chasing ghosts; The symptoms are as follows: I'm have an excellent ADSL connection, but connecting to certain servers using timeout-sensitive protocols I am having problems. For example, when trying to upload files to my FTP server, either using FTP or SSH + rz, the connection takes forever and breaks up in the middle quite frequently. Pinging the server shows that my packet loss is negligible and that the connection is fast (~35ms, 1% packet loss). Other people can FTP with no problems. I have no other problems with that server or with my Internet connection in general (i.e SMTP, HTTP all work quite nicely). The only think I can think of is some strange MTU problem. that sounds likely, according to the symptoms you describe. does this happen when connecting through the adsl masquerading server (the computer which runs pptp) or only when connecting through a masqueraded client? For example, trying to FTP from my linux connection (the one connected to an ADSL) via FTP fails miserably with timeouts. The connection is done directly, so it's not a masquerading problem. Now the facts: The MTU on the ppp0 interface is: 1452 The MTU on the eth1 interface (the one connected to the ADSL modem) is: 1500 sounds correct. as far as I can tell from the how-to, that should be the right values. Any idea how I can debug it and/or fix the problem? use the ethereal, luke :) -- The ill-formed Orange Fails to satisfy the eye: http://vipe.technion.ac.il/~mulix/ Segmentation fault. http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: how mature is openoffice ?
On Fri, Mar 01, 2002 at 10:58:12AM +0200, Nadav Har'El wrote: On Thu, Feb 28, 2002, mulix wrote about Re: how mature is openoffice ?: ObSignal: any advice for a person who hates writing html but wants a convenient way to create a personal website? Read up on CSS (cascasding style sheets, w3.org has an excellent document about it), and do most of the visual stuff in CSS, not HTML, leaving you with only a minimal amount of HTML you'll need to use. since my pages have almost no visual content, that would be a rather moot point :) [of course, the more straighforward answer to your question is use Mozilla's editor, or something like that...] not surprisingly at all, that's what i did at first, until i discoverd how buggy it is and the awful html it produced (hurts my eyes to look at it). now i do it in xemacs, with a couple of python scripts to generate page templates and to checking the changes into cvs, commit and upload to the remote site. scripts are available on the site if anyone's interested, but they're ugly. You Have Been Warned. -- The ill-formed Orange Fails to satisfy the eye: http://vipe.technion.ac.il/~mulix/ Segmentation fault. http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: how mature is openoffice ?
On Thu, Feb 28, 2002 at 12:07:55PM +0200, Erez Doron wrote: how mature is openoffice ? why dont *you* try it and tell *us*? not to mention that it's probably been discussed to death on this list already. i usually ignore all such threads, so i wouldn't really know. ObSignal: any advice for a person who hates writing html but wants a convenient way to create a personal website? -- The ill-formed Orange Fails to satisfy the eye: http://vipe.technion.ac.il/~mulix/ Segmentation fault. http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: personal web site [was: Re: how mature is openoffice ?]
On Thu, Feb 28, 2002 at 02:43:27PM +0200, Tzafrir Cohen wrote: On Thu, 28 Feb 2002, mulix wrote: On Thu, Feb 28, 2002 at 12:07:55PM +0200, Erez Doron wrote: how mature is openoffice ? why dont *you* try it and tell *us*? not to mention that it's probably been discussed to death on this list already. i usually ignore all such threads, so i wouldn't really know. ObSignal: any advice for a person who hates writing html but wants a convenient way to create a personal website? Is maintaining the page from a remote shell a requirement? not sure what you mean here. the page (several pages, actually) is hosted on a remote machine and is updated via the 'tar; scp; untar' method. -- The ill-formed Orange Fails to satisfy the eye: http://vipe.technion.ac.il/~mulix/ Segmentation fault. http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
2.4.18 available on http://www.iglu.org.il/lxr
lxr, which is linux cross reference, is a nifty tool which helps kernel hackers by providing extensive cross references within the kernel sources. shlomif installed lxr on http://www.iglu.org.il/lxr, and i've been maintaining it lately. i plan to add each new 2.4.x kernel release (-final, not -pre or -rc), and 2.4.18 is available now (has been available for since yesterday, but i forgot to announce it). if anybody wants any other kernel trees (currently we have 2.4.9, 2.4.17, 2.4.18 and 2.5.3), feel free to ask. -- The ill-formed Orange Fails to satisfy the eye: http://vipe.technion.ac.il/~mulix/ Segmentation fault. http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: 2.4.18 available on http://www.iglu.org.il/lxr
On Wed, Feb 27, 2002 at 05:31:15PM +0200, Oleg Goldshmidt wrote: mulix [EMAIL PROTECTED] writes: if anybody wants any other kernel trees (currently we have 2.4.9, 2.4.17, 2.4.18 and 2.5.3), feel free to ask. 2.2(.19) would be nice if it is not a big trouble. Not critical for me but might come handy. 2.2.20 is the latest published, i'm if'm not mistaken. it's not a problem at all - which would you prefer? -- The ill-formed Orange Fails to satisfy the eye: http://vipe.technion.ac.il/~mulix/ Segmentation fault. http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: 2.4.18 available on http://www.iglu.org.il/lxr
On Wed, Feb 27, 2002 at 07:15:32PM +0200, Shaul Karl wrote: On Wed, Feb 27, 2002 at 05:31:15PM +0200, Oleg Goldshmidt wrote: mulix [EMAIL PROTECTED] writes: if anybody wants any other kernel trees (currently we have 2.4.9, 2.4.17, 2.4.18 and 2.5.3), feel free to ask. 2.2(.19) would be nice if it is not a big trouble. Not critical for me but might come handy. 2.2.20 is the latest published, i'm if'm not mistaken. it's not a problem at all - which would you prefer? IIRC 2.2.19 has a security hole (local root exploit of some sort?). not relevant in the context of lxr. 2.2.20 is added to the lxr. the indexing script chocked on a few files in fs/nls/cp*, so those are missing. -- The ill-formed Orange Fails to satisfy the eye: http://vipe.technion.ac.il/~mulix/ Segmentation fault. http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
mua which handles very large mailboxes
hello, linuxers, anyone knows a good mailer, command line based, which can handle very large mailboxes? on the order of thousands and tens of thousands of messages (think lkml archive). mailers i'm not interested in: pine (i'm using it right now, doesnt cut it above several hundred messages), evolution, kmail, mozilla, any other x based mailer. i read my mail over ssh frequently, and an x mua is not feasible. i'll be checking out mutt very shortly. anything else i should look into? -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: mua which handles very large mailboxes
On Sun, Feb 24, 2002 at 12:24:15PM +0200, Christoph Bugel wrote: On Sun 2002-02-24, mulix wrote: hello, linuxers, anyone knows a good mailer, command line based, which can handle very large mailboxes? on the order of thousands and tens of thousands of messages (think lkml archive). I'm using mutt. mutt supports many mailbox formats, but I use the default mbox one. I know someone who uses a mua where every message is a file. it's called gmh or mh I think. if you'll look at my headers, you'll see that i'm using mutt as well as of now. its handling of large mailboxes does leave something to be desired, but it's definitely better than pine. i've also been looking for a good reason to ditch pine, due to it's non-free license. nadav - thanks for your csplit recommendation. however, i'm interested in an mua which will let me view *one* huge lkml folder at once. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: mua which handles very large mailboxes
On Sun, Feb 24, 2002 at 12:40:46PM +0200, Oleg Goldshmidt wrote: mulix [EMAIL PROTECTED] writes: anyone knows a good mailer, command line based, which can handle very large mailboxes? on the order of thousands and tens of thousands of messages (think lkml archive). What do you mean by handle? Like actually reading mail using it? I use GNUS, which can handle large amounts of mail as it is originally a newsreader. It is not exactly command-line (emacs), and implies a bit of a learning curve (probably no deterrent for you). gnus is indeed a very good option, since i already spend most of my time inside (x)emacs's many buffers. i'll check it out once i get mutt working to my satisfaction. which leads me to the last remaining problem with mutt - i want it to show new mail which arrives to the spool file *immediately* when it arrives. right now, it only shows it when i press a key. rtfm'ing has been less than useful, and i have 'set mail_check=1' in my .muttrc. ideas? -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: mua which handles very large mailboxes
On Sun, Feb 24, 2002 at 03:20:37PM +0200, Geoffrey S. Mendelson wrote: mulix wrote: mailers i'm not interested in: pine (i'm using it right now, doesnt cut it above several hundred messages), evolution, kmail, mozilla, any other x based mailer. i read my mail over ssh frequently, and an x mua is not feasible. Have you looked into IMAP? IMAP is a protocol that keeps mail on the server. Your IMAP client interogates the server asking it if there is new mail, sends that mail t the client to be read, and stores any replies, etc on the server. since i run my own mail server, why would i want to use IMAP? i find it much easier to ssh into my box from wherever i'm at (using the wonderful putty ssh java applet if no trusted ssh client is installed on the local computer) and read mail as normal user. Sounds like POP? Well IMAP is the son of POP, the biggest difference to the user is that mail is stored in one central place. Sounds like it's exactly what you need. imap is also *plain text*. need i say any more? as for my problem with mutt, it's solved. there's a small configuration variable called 'timeout', with controls how long mutt will wait for a user to be inactive before it will check new mail. setting this variable to something sane, like 10 seconds, instead of the default 600, solved the problem. mulix, a mail junky. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: mua which handles very large mailboxes
On Sun, Feb 24, 2002 at 04:16:10PM +0200, Orr Dunkelman wrote: On Sun, 24 Feb 2002, Nadav Har'El wrote: On Sun, Feb 24, 2002, mulix wrote about Re: mua which handles very large mailboxes: imap is also *plain text*. need i say any more? Ah? What do you mean imap is plain text? IMAP is just a protocol for remote access to mail messages (IMAP=Internet Message Access Protocol). These mail messages might contain pictures/sounds/ video/whatever (i.e., MIME), and you can view them on a graphical client (e.g., Mozilla which has excellent IMAP support). I believe that mulix talked about encryption (plain text = not encrypted). Thus, all the messages can be understood by the entire network. close, but not entirely correct - i was referring to the autentication phase, where your user name and password are sent in clear text. i guess i should've said more... -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: mua which handles very large mailboxes
On Sun, Feb 24, 2002 at 05:34:11PM +0200, Tzafrir Cohen wrote: On Sun, 24 Feb 2002, mulix wrote: On Sun, Feb 24, 2002 at 12:24:15PM +0200, Christoph Bugel wrote: On Sun 2002-02-24, mulix wrote: hello, linuxers, anyone knows a good mailer, command line based, which can handle very large mailboxes? on the order of thousands andtens of thousands of messages (think lkml archive). I'm using mutt. mutt supports many mailbox formats, but I use the default mbox one. I know someone who uses a mua where every message is a file. it's called gmh or mh I think. if you'll look at my headers, you'll see that i'm using mutt as well as of now. its handling of large mailboxes does leave something to be desired, but it's definitely better than pine. i've also been looking for a good reason to ditch pine, due to it's non-free license. You've got to be some purist to have issues with pine's license. even if we ignore all of the other issue with pine, it's not free software (according to debian's guidelines). why should i run it on mystem when there are better free alternatives? for the curious, http://www.asty.org/articles/20010702pine.html -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: gcc upgrade
On Thu, 14 Feb 2002 [EMAIL PROTECTED] wrote: We're developing on stock RH 6.2 (for product compliency), but ran into a gcc bug when overriding new/delete operators. It was fixed in later versions of gcc (2.95.2 +) , so now we want to upgrade compiler. Question: What does upping to gcc 3.x involve (other than installing an RPM), what could be the ramifications on how our (quite big) project compiles, and how well is its backwards-compatibility with gcc 2.x? Should I just take 2.95.3 if I want to sleep well at night? you didnt mention if your code is c or c++, how big it is, if it has strict standards conformance or if it's kernel code, all of which could influence the answer to your question. in general, gcc3 is much more conforming to the c++ standard, and breaks binary compatibility wrt c++ ABI with earlier versions. it's also relatively new, compilerwise, so it might still have some bugs. gcc295, on the other hand is a maintenance release, and is pretty stable and close to the compiler you're using. basically, upgrading the compiler is no big deal, and several compilers can co-exist peacefully (my systems have three compilers on them at the moment). just get a test box, install the new compiler on it and try... if you're installnig from rpm you might have to play with the --prefix or some such, unless the compiler is meant to be installed alongside another compiler (like gcc, kgcc and gcc3 on rh7.2 systems). if installing from source, *read the documentation*, although it all boils down to a few options to configure, if i remember correctly. hope this helps, more info would probably help us help you. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: [SUMMARY]:book recomendation needed
On Sun, 10 Feb 2002, Shlomi Fish wrote: Professional does not mean experienced. Naturally, I don't expect someone who read such a book to be as good as someone who programmed it for many years. But, of course, some people who programmed C (for example) for 1 year write much better code that those who programmed it for 5 years. Time the other way around, dont you think? is not the only factor that determines the quality of the programmer. of course not, but i dont consider someone who 'read the book' a professional either, even if it's a really good book. fwiw, my personal opinion is that a professional is someone who a) knows what he's doing b) has the code to prove it ObLinux: looking for a good, secure, free (speech and beer), simple and hopefully small enough to audit by hand ftp server. suggestions? Sorry, cannot help you as I did not have any experience with experiencing with my own ftp server. I believe I have an FTP server on my Com-Net workstation, but it is what came with the MDK 8.1 distro. You do have quite a lot of demands, though. i'll probably end up writing it on my own, in python, since i only need a subset of the functionality. that's what i did when i needed a web server, except i used java at the time. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: [SUMMARY]:book recomendation needed
On Sat, 9 Feb 2002, b g wrote: Tom Swan's C++ and Thinking in C++ by B.Eckel (great book!!!) that one (tic++) is available on line, along with thinking in java and thinking in patterns, and maybe also thinking in python (been awhile since i browsed eckel's website). excellent books indeed. on a general note, i always found computer books with titles such as these: gar unleashed using foo teach yourself cobol in 24 seconds complete guide to spamming foobar for morongs sorely lacking. on the other hand, anything by adison wesley, oreilly or prentic hall is usually a very good buy. ymmv. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: Basic Compiler
On Sat, 9 Feb 2002, U. P. wrote: Hey ! Iam looking for a Basic compiler for linux, any ideas where can i get one ? gcc.gnu.org, although you will probably better off getting the g** rpm and all of its dependencies from your distribution. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: Basic Compiler
On Sat, 9 Feb 2002, mulix wrote: On Sat, 9 Feb 2002, U. P. wrote: Hey ! Iam looking for a Basic compiler for linux, any ideas where can i get one ? gcc.gnu.org, although you will probably better off getting the g** rpm and all of its dependencies from your distribution. whoops, apologies, i misparsed Basic as basic. why on earth would someone want a Basic compiler, anyway? -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: [SUMMARY]:book recomendation needed
On Sat, 9 Feb 2002, Shlomi Fish wrote: On Sat, 9 Feb 2002, mulix wrote: sorely lacking. on the other hand, anything by adison wesley, oreilly or prentic hall is usually a very good buy. ymmv. I would not say that anything by O'Reilly, etc. is a very good buy. For instance, I don't see the point of buying an entire book just to learn Sed and Awk. And I bet that it would be redundant to buy some of their Perl books. Don't get me wrong, their books are usually very professional and all, but they are sometimes too specific. then how would you suggest one learn sed awk? note - when i say learn i really do mean learn. note write one script based on the examples and call it quits As for the Unleashed, Teach yourself, etc. Those books are obviously intended for a less professional crowd who wishes to become familiar with a given technology as quickly as possible, while being made aware of all the caveats it contains. Some of them are actually pretty good, although expert hackers may find them too slow-paced. (how to create a button... how to create a listbox... how to create a combo-box...) i have read several unleashed/24/using books, and all of them were either incomplete, shallow or plain wrong. someone who *thinks* he knows something but is wrong is a lot more dangerous than someone who simply doesn't know. My problem is that I have an on-demand way of learning something new. What I mean is that I use a sub-set of the technology and when I need more, or feel that something is missing, I learn it by looking for info on the web. That's not the best way of mastering something, but I seem to like it. Besides, I'm almost sure nobody uses the whole of C++, Perl, Common Lisp, etc. Those languages have so much redundency over Turing Completeness, that using a subset will not hurt too much. ;-) you need to differentiate between learning enough to get the job done and becoming a guru on the subject. when talking about book learning, i usually refer to the second definition. for the first, the web/manuals/examples/source code is almost always enough, if you have the underlying theory. [followups should probably go to hackers-il?] -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: [SUMMARY]:book recomendation needed
On Sun, 10 Feb 2002, Shlomi Fish wrote: You cannot expect Java Unleashed (which I did not thoroughly read, but found interesting and was quite impressed from), to teach you _everything_ there is to know about Java. Java is a very encompassing technology and its getting worse in this sense. But it can bring you to a stage where you can advance on your own, and be familiar with all the important caveats of it. that has not been my experience with the unleashed books, but the last one i bothered to read was several years ago, so maybe they got better since. (doubtful, but possible). Well, so do I. However, I think that such books can teach you enough to become a professional, although not very experienced programmer in a certain technology. you seem have a curious definition of 'professional'. ObLinux: looking for a good, secure, free (speech and beer), simple and hopefully small enough to audit by hand ftp server. suggestions? -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: The Great Kernel CVS Mutiny
On Thu, 7 Feb 2002, Ely Levy wrote: there is no kernel of any OS that I know that only one person decide usualy there are few people and a voting involved. designed by a comittee is *not* a compliment. not mention that not EVERY patch goes to the that person neither should every patch go to linus. see recent patch penguin thread on lkml. /me vows to stay out of pointless, meaningless, gossip driven, fact lacking threads for at least 24 hours. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
jabber CLI client and X over ssh
hi guys, anyone got a good jabber CLI client to recommand? also, anyone knows if it's possible to run an X program over ssh, when you're ssh'ing through more than one host? desktop --- gateway --- remote computer and have a remote computer X program display on desktop. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: strange tcsh problems
On 4 Feb 2002, Erez Doron wrote: On Sun, 2002-02-03 at 14:50, mulix wrote: On 3 Feb 2002, Erez Doron wrote: what can you say about this: bash-2.03# /usr/src/build/tcsh-6.11.00/tcsh [erez@familiar ~]$ ls ls: Command not found. [erez@familiar ~]$ /bin/ls Applications Settings bin inews mp3 pymote [erez@familiar ~]$ echo $PATH /bin:/sbin:/usr/bin:/usr/sbin:/usr/X11R6/bin:/usr/local/bin:/usr/local/sbin:/root/bin:/usr/bin:/bin:/usr/sbin:/sbin [erez@familiar ~]$ echo $path /bin /sbin /usr/bin /usr/sbin /usr/X11R6/bin /usr/local/bin /usr/local/sbin /root/bin /usr/bin /bin /usr/sbin /sbin [erez@familiar ~]$ [erez@familiar ~]$ rehash [erez@familiar ~]$ ls ls: Command not found. [erez@familiar ~]$ alias [erez@familiar ~]$ alias ls [erez@familiar ~]$ which ls ls: Command not found. [erez@familiar ~]$ now you've got me intrigued. complete 'strace ls' output, please. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: kapm-idled
On Sun, 3 Feb 2002, Tal Amir wrote: [/me boggles at the abuse infected on pure kapm-idled]. any1 knows how to kill this SOB ? ;) dear tal, kapm-idled is a *kernel* thread. that means, in the nicest possible way, Do Not Fuck With It. a quick google search points one to http://www.google.co.il/search?q=cache:1vbuMuhuMeEC:kt.linuxcare.com/kernel-traffic/kt20010101_100.epl+kapm+kthl=iw you want part 4. [long and probably warped url of the google cache of the kernel traffic story]. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: strange tcsh problems
On 3 Feb 2002, Erez Doron wrote: what can you say about this: bash-2.03# /usr/src/build/tcsh-6.11.00/tcsh [erez@familiar ~]$ ls ls: Command not found. [erez@familiar ~]$ /bin/ls Applications Settings bin inews mp3 pymote [erez@familiar ~]$ echo $PATH /bin:/sbin:/usr/bin:/usr/sbin:/usr/X11R6/bin:/usr/local/bin:/usr/local/sbin:/root/bin:/usr/bin:/bin:/usr/sbin:/sbin [erez@familiar ~]$ echo $path /bin /sbin /usr/bin /usr/sbin /usr/X11R6/bin /usr/local/bin /usr/local/sbin /root/bin /usr/bin /bin /usr/sbin /sbin [erez@familiar ~]$ which ls; alias ls; those two might provide a clue. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: Help compiling mpeglib_divxplugin
On Wed, 30 Jan 2002, Geoffrey S. Mendelson wrote: Erez Boym wrote: Hi, I'm trying to view DivX files on my Linux box (Mandrake 8.0, KDE 2.1.1, Nuaton 1.0.1). Try mplayer, instead. plays a lot more files, though it won't play VCD's. i second that, mplayer works great for me. and while you compile it, read their website, especially the gcc-2.96 is a devil spawn bits. quite hilarious, really. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: Kernel: Behaviour of kmalloc in AA's new VM
On Thu, 24 Jan 2002, Shlomi Fish wrote: We noticed our kernel module OOPSified kernels that used the new VM by Andrea Archanageli, and we assumed it was still buggy, so we sticked with the older one (which is by Rik van Riel IIRC). Now, in the last IGLU meeting I had a talk with Moshe Bar who said, that it was not supposed to work. The reason is that we used kmalloc and if we don't touch the buffer we allocated for a long time, and then tried to access a pointer to a 5th or greater page after its start, it is not guaranteed to succeed. that sounds completely bogus. i presume there was a misunderstanding somewhere. rant kmalloc is supposed to be similar to the malloc() userland function, which gives you a dynamically allocated buffer for life. Why does not kmalloc behaves that way in the new VM? Would it not break compatibility with many third-party modules, like our own? /rant i dont believe it acts this way. fwiw, are you allcoating memory *only* using kmalloc? where does your code oops? Can anybody clarify this situation? sure, but probably not on this list. you might want to try either kernelnewbies or the linux kernl maling list. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: AOL in Negotiations to Buy Red Hat? (fwd)
bravo. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ -- Forwarded message -- Date: Mon, 21 Jan 2002 16:35:45 + (GMT) From: Alan Cox [EMAIL PROTECTED] To: M A [EMAIL PROTECTED] Cc: robert w hall [EMAIL PROTECTED], [EMAIL PROTECTED] Subject: Re: AOL in Negotiations to Buy Red Hat? well - AOL could replace that nice-looking young lady in their tele-ads (the one with the dress like a vdu) with Alan Cox in Wizard's hat n cloak :-) -- robert w hall Well I've no idea on the rumours (and if I did I wouldnt tell you!) but Im insulted that anyone believes I would continue working for RH if aol/time warner owned them. Alan - To unsubscribe from this list: send the line unsubscribe linux-kernel in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: The Great Kernel CVS Mutiny
[/me replies against my better judgement. oh well] On Sat, 19 Jan 2002, Shlomi Fish wrote: I believe that the way things are done in the kernel development have to change in many ways to make its maintainance more scalable and straightforward. There should be: why should its maintanence be more scalable and straightforward? scalability is a nice buzzword, but so is quality, cohesion, direction, which linus supplies in the current system. 1. CVS. why? you haven't given *one* compelling argument why kernel hackers need a cvs. as adi aptly put it, linus is a cvs with taste. no source control system will replace him. 2. Roadmaps for the next stable release - what goes in, what stays out, etc. sometimes there is. it's posted on linux-kernel periodically by linus, usually in a rather crude form. (bio in, kbuild not yet, etc). 3. A general, short, manifesto that explains what Linus thinks the kernel should have and what not. (I.e: should it have a sound subsystem? Should it have GGI? Should it have an in-kernel GUI?). Just kidding about the GUI part. i think it's pretty obvious what should be in a unixish kernel and what shouldn't be. 4. Maintainers of different subsystems who may delegate authority to other people. we have that. 5. Making sure Linus need not be bothered with any little patch. Linus can read the CVS diffs if he'd like, but people can commit changes without asking him. linus WANTS to be bothered with any little patch. it's a crucial part of his quality control system. 6. Defining well-defined interfaces for the subsystems to interact with each other, and explaining what should or should not be done. this is called a micro kernel, unlike linux, which is a monolithic kernel. Did I miss anything? I believe what I proposed is better than the ways things are done now. yes, you missed quite a lot. for starters, explain why your system will be better than the system we have now, which is not perfect, but works. i also get the feeling you are not very familiar with the way linux kernel development is done, except from second and third hand sources. may i suggest you take some time to *learn how its done now*, before proposing a Grand Novel Absolutely Better Way of Doing Things? also, this subject has been hashed to death on lkml. maybe you should read an archive or three? -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: turn off those ppp0-n
On Thu, 17 Jan 2002, Nx wrote: first i wanna 10Q 4 helping me with connecting my ADSL 2 the net now, i after i turn it off(the progs:pppd,pptp) i cant get it back on why not? what is the error you receive? please be verbose and precise in your answer. and i get in the ifconfig -a ppp0,ppp1 and so on are they marked UP or not? PS how do i secure my box better? google is your friend. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
FAQ: how to ask smart questions
this has probably been posted here before, but based on the recent threads, it surely bears repeating: How to Ask Smart Questions: http://tuxedo.org/~esr/faqs/smart-questions.html quoting: ...hackers have a reputation for meeting simple questions with what looks like hostility or arrogance. It sometimes looks like we're reflexively rude to newbies and the ignorant. But this isn't really true. What we are, unapologetically, is hostile to people who seem to be unwilling to think or do their own homework before asking questions. People like that are time sinks -- they take without giving back, they waste time we could have spent on another question more interesting and another person more worthy of an answer. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: computer hangs (was RE: The Power of Open-Source)
On Thu, 17 Jan 2002, Daniel Feiglin wrote: For what it's worth, I have the kscd applet running (CD player), and quite frequently before such a hang, the applet crashes (SIGSEGV). However that is not exclusive. I smell something in the KDE panel - perhaps one or more of the applets. The trouble is, that in a hang of this kind, nothing gets to the logs. let me get this straight - it's KDE that's crashing, and perhaphs also X, right? why do you think nothing gets to the log? run kde under strace. enable whatever debug options kde has. and the next time this happens, ssh into your machine from another machine, and see what you can gleam from the logs. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: computer hangs (was RE: The Power of Open-Source)
On Thu, 17 Jan 2002, Daniel Feiglin wrote: let me get this straight - it's KDE that's crashing, and perhaphs also X, right? Can't say for sure. The only way to establish that would be to run a different Window Manager for an extended period. i meant that it's not the kernel crashing - in other words, it's not a complete lockup. why do you think nothing gets to the log? 'Coz I looked!! i've yet to see a project that doesnt have an option to turn on debugging output. i dont doubt the fact you looked, but i highly doubt the fact that under no circumstances does anything get written to the logs. as a last resort, run kde under strace [or syscalltrack, when it's mature enough ;)] -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: Stable kernel update
On Tue, 15 Jan 2002, Hetz Ben Hamo wrote: Kernel 2.4.17 + Marcello's pre3 + Alan Cox ac2 patch = kernel 2.4.18-pre3-ac2 is the the most stable tests with the new VM, after 12 hours of heavy stress testing (tests include: parallel compiling X + KDE + running OpenGL demos + live file system backup).. i find your use of 2.4.18pre3-ac2 and 'stable' in the same sentance amusing, if a bit oxymoronic, since that kernel has been out for all of two days. maybe it does beautifully under stress, but what if it has a hidden bug that causes it to degrade over time, so that in a week it's unusable? [i doubt it, i'm running 2.4.18pre3-ac1 myself]. So, if you want to compile it - then: 1. grab kernel 2.4.17 2. grab kernel 2.4.18pre3 patches from v2.4/testings/patch-2.4.18-pre3.tar.gz 3. grab Alan's AC2 patch find it at ftp://ftp.kernel.org/pub/mirrors/kernel.org/linux/kernel/people/alan/linux-2.4/2.4.18/ Apply them all (first open the 2.4.17, then gzip -d apply Marcello's patch, then gzip apply Alan's patch).. no need to gzip. zcat does the job fine. newcomers to patch might want to pass to it the --dry-run option first, to verify the patch will apply cleanly. The above kernel got also support for 130+ GB hard drives (like Maxtor's new ones) that would be andre hedrick's ide stuff. it also contains rik van riel and friends' latest vm work, rmap. read Alan's ChangeLog if you want to know more. hetz, i'm glad this kernel is so stable for you. did you let lkml know? or at least the developers themselves? /me, who should stop reading lkml this mornign and start writing code. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: Stable kernel update
On Tue, 15 Jan 2002, Hetz Ben Hamo wrote: I'm rather confused at the moment. Does this kernel contain the new Virtual Memory manager from Andrea Archangeli or does it use the original VM that was present in the older 2.4.x releases? The new one. Old one is deprecated officially (look at Redhat's Rawhide kernel - 2.4.16) new one indeed, but heavily modified and extended by rik van riel, author of the original vm. you can find his patch posted periodically to the lkml mailing list under the heading 'rmap' [short for reverse mapping - mapping from pages in memory to the process that owns them]. here's what his latest announcement said: From [EMAIL PROTECTED] Tue Jan 15 09:36:57 2002 Date: Fri, 11 Jan 2002 14:54:08 -0200 (BRST) From: Rik van Riel [EMAIL PROTECTED] To: [EMAIL PROTECTED] Cc: [EMAIL PROTECTED] Subject: [PATCH *] rmap VM #11b The second maintenance release of the 11th version of the reverse mapping based VM is now available. This is an attempt at making a more robust and flexible VM subsystem, while cleaning up a lot of code at the same time. The patch is available from: http://surriel.com/patches/2.4/2.4.17-rmap-11b andhttp://linuxvm.bkbits.net/ I know that I could not get the IP-Noise kernel module to run on the AA VM, while it ran on the old VM flawlessly. Eww, donno... to me that would seem to point to bugs in your code, shlomi. as far as i understnad, IP-Noise should have no business with the vm, only using its exported interfaces. what exactly where the problems? [offlist if you'd like]. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Bug in alcatel speed touch home adsl modem (fwd)
greetings. the following (unsubstantiated) mail from bugtraq might interest alcatel speed touch home modem users. could anyone who used the 'expert' hack please let me know if the fact that the modem does masquerading means that it has an externally visible ip? -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ -- Forwarded message -- Date: Fri, 11 Jan 2002 18:52:04 +0100 From: Hacknisty [EMAIL PROTECTED] To: [EMAIL PROTECTED] Subject: Bug in alcatel speed touch home adsl modem Hi I've found a bug in Alcatel Speed Touch Home ADSL modem I was playing with nmap when i've tried to scan my modem with is local ip address (default is 10.0.0.138 but i've changed it to 10.0.0.1) I've tested various scan and note that when you activate OS detection, modem reboot immediately Is anyone can confirm that ? My firmware version is 8706 I don't think we can do anything with that but i want to know what really happened on the modem. Is anyone can explain that to me Sorry for my poor english and sorry to if this was already post here Amically Hacknisty = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: Small C question
On Thu, 10 Jan 2002, Daniel Feiglin wrote: Several well known utilities like dd, fdformat use a function _(...) in the error reporting calls. What does it do and where is it documented? (Checked GNU V Library docs, cc docs - nothing doing.) it's an application specific convention, so you wouldn't find it documented anywher. ususally used to wrap the actualy function with a macro which willa automatically supplies the file name, function name and line number. consider: [example snarfed from the pptp sources] void _fatal(char *func, char *file, int line, char *format, ...) __attribute__ ((format (printf, 4, 5))) __attribute__ ((noreturn)); #define fatal(format, args...) \ _fatal(__FUNCTION__,__FILE__,__LINE__, format , ## args) this code uses two gcc specific extensions, but i'm sure you get the point. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: Small C question
On 10 Jan 2002, Alex Shnitman wrote: On Thu, 2002-01-10 at 17:34, Daniel Feiglin wrote: Several well known utilities like dd, fdformat use a function _(...) in the error reporting calls. What does it do and where is it documented? (Checked GNU V Library docs, cc docs - nothing doing.) It enables the program to output locale-dependent text. The GNU gettext suite works with this macro. It has a define: #define _(String) gettext(String) It also has a tool that goes over your source files, extracts all the strings that are wrapped this way, writes them all out to a .po file and lets you create translations for all the text strings in your program. Then if everything is configured properly, they will automatically be loaded and used at runtime depending on the LANG environment variable. (The gettext function takes care of that.) whoops. this is indeed the correct answer. i completely misparsed the original question. sorry if i confused anyone with the correct answer, albeit to a completely different question :) -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: BEZEQ_ISRAEL
On Wed, 2 Jan 2002, Eli Marmor wrote: I use mulix' PPTP under an embedded Linux, but want to move to the which distro? just curious. standard PPTP of the distro, which doesn't support the patch, even not through quirks. that's because there hasn't been a pptp release since the quirks patch was commited. From first look, everything seems to work correctly, even with the standard PPTP; Is it possible? Is the patch needed only for a certainly possible. specific ADSL modem? (I use Alcatel, and I remember discussions about ATUR2 vs. ATUR3 of Orckit; So among the three - ATUR2/ATUR3/Alcatel - which needs this patch and which doesn't?) alcatel works fine with and without the patch. atur 2 requires some of the patch atur 3 requires all of the patch. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: BEZEQ_ISRAEL
On Wed, 2 Jan 2002, Eli Marmor wrote: Dani Arbel wrote: Alcatel ethernet modem needs no patch. Thanks! I understand that the problematic modem is Orckit-Atur3, while Atur2 (and Alcatel of course) needs no patch. Am I wrong? slightly, atur 2 needs a small patch as well. I hope that there will be a new (official) version of PPTP soon. Dani/ mulix: Does any of you have any idea about PPTP developers plans? the pptp maintainer intends to make a release Real Soon Now(tm). the problem is that he wants/needs to make rpms, and none of us know how to do that. the previous maintainer promised to help with that, but it hasn't happened yet. when it does, we'll have new release, with the quirks support. if anyone has experience with making rpms, and would like to help, let me know. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: firewall script
On Tue, 1 Jan 2002, Tzafrir Cohen wrote: I want to convert my firewall from kernel 2.2 to kernel 2.4 . I believe that 2.4 is by now stable enough. make sure to go straight to 2.4.17. pretty much every other version had known problems. My main limitation with this system is that I would like to minimize the console time spent near it. Furthermore, I don't have much of a testing environment, so I would like to start with s script that is generally know to work, and has all the major features that I need. When browsing over project lists in freshmeat I can see features list, but stability is not something aparent from there. moast such scripts simply call iptables. where does stablity come into play? I would also prefer a system that does some sanity-checking to the rules before applying them (to minimize the chance of locking myself out because of a simple typo). Major features that I need: * NAT check * DMZ donno * Forwarding of internal ports check. Any recomendations? i use a heavily modified version of monmotha's firewall. it's easy to understand and modify, and does the job for my lan. http://monmotha.mplug.org/firewall/index.php [looking at the site, i see there were some *security* problems with it lately (although i was not affected). so take this recommendation with a grain of salt]. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: firewall script
On Tue, 1 Jan 2002, Tzafrir Cohen wrote: On Tue, 1 Jan 2002, mulix wrote: On Tue, 1 Jan 2002, Tzafrir Cohen wrote: I want to convert my firewall from kernel 2.2 to kernel 2.4 . I believe that 2.4 is by now stable enough. make sure to go straight to 2.4.17. pretty much every other version had known problems. Which means I have to wait until 2.4.18, because 2.4.17's problems are currently yet unknown ;-) surely you see the recursive problem here :) Speaking about that: the latest kernel source from Mandrake cooker (kernel-source-2.4.16.11mdk-1-1mdk) fails to compile for me, with some error in loop.c i thought mandrake (and other distros) kernels are to be rpm -Uvh'd or such. any special reason for compiling it? (except from that warm glowing feeling we all get from hunt^H^H^H^H compiling our kernels? My main limitation with this system is that I would like to minimize the console time spent near it. Furthermore, I don't have much of a testing environment, so I would like to start with s script that is generally know to work, and has all the major features that I need. When browsing over project lists in freshmeat I can see features list, but stability is not something aparent from there. moast such scripts simply call iptables. where does stablity come into play? I would also prefer a system that does some sanity-checking to the rules before applying them (to minimize the chance of locking myself out because of a simple typo). To answer the previous question: If I have to edit the script for every configuration change (e.g: open /close a port) then I run into a risk of creating a syntax error that will leave the system in ahalf-configured situation (think of an extra ') ditto if the script has a seperate config file, but sources it, rather than parsing it. gotcha. good point, especially if working remotely. i use a heavily modified version of monmotha's firewall. it's easy to understand and modify, and does the job for my lan. http://monmotha.mplug.org/firewall/index.php I'll have a look Though this seems to be lack a start and stop comands of a standard sysv-init script . alias fw-start=/path/to/script alias fw-stop=iptables -F -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
ANN: adsl-howto 2.2.2 (minor version) released
a new minor version of the adsl howto has been released. from now on, the adsl howto will be available from: http://damyen.technion.ac.il/~dani/adsl-howto.txt [same as always] http://vipe.technion.ac.il/~mulix/adsl-howto.txt [this one changed] the only new thing in this version is that i updated various urls which pointed to www.pointer.co.il to point to vipe.technion.ac.il. [1] due to reception of information for both freebsd and openbsd and restriction of the text format, we intend to convert and extend the howto to an LDP style format. people who want to help with the rewrite, please contact dani arbel or me. [1] i'd like to thank pointer software systems for having hosted my homepage and the howto until now. alas, i no longer work there as of tomorrow night, so my homepage, including the howto and the pptp patches, have moved to http://vipe.technion.ac.il/~mulix/ -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: Earlier this week: Linux Confrence at compaq
On Sat, 29 Dec 2001, Orna Agmon wrote: i was there, and so was mulix, as you can see in http://www.advogato.com/person/mulix yup, i was. i see now that i sent my impressions to shaul karl in private, so i shall add them here in case the list finds them interesting. On Fri, 28 Dec 2001, Shaul Karl wrote: IGLU front page has an announcement about `Linux Confrence at compaq' that was held earlier this week. Can someone who attended post a short summary: * many / few people attended * worth / not worth the time and so on. More details then a short summary is desiarble. here are my *personal* impressions: compaq (as represented by the people lecturing) does not understand linux too well. example - they wrote that freebsd is a type of linux distribution. nonetheless, they see it as a viable market and support it fully. compaq also supports linux and open source - iglu was mentioned once or twice, as well as www.opensource.compaq.com. the lectures were very high level, as could be expected, and mostly marketish in nature. the dealt specifically with clustering on linux, not just linux. the audience didn't understand linux/open source/free software too well, in general. people continously asked and you can download it from the internet, just like that?. the audience was mostly mature, not many people i would consider programmers (from appearance only!) but rather IT managers, etc. all in all, a very interesting day, much more interested than i expected it to be. even the one boring lecture was productive, since i coded a much needed feature for syscalltrack during it :) anyone interested in specific aspects presented or not presented, is welcome to ask me in private. likewise. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: Care and Feeding of Stupids (was: Re: stupid me)
[while i avoided the first incarnation of this discussion, it's been a rather dull day so far, and i'm aching for excitement. flamefest about to erupt - dont keep reading if you dont want to get burned.] On Fri, 28 Dec 2001, Nimrod Simba Carmi wrote: A while ago, I posted here a message looking for programmers to help us build a content engine system for School Sucks. quite the stupidest name, imho, but never mind that. I got tens of flames about it, people were amazed how come I dared asking for programmers without willing to pay more then 20$/hour. why would people care how much you're willing to pay? if they think it's too low (which i think it is, by far, but then i wouldn't consider this type of job for a myriad of other reasons) they simply shouldn't take it. Well I just wanted to say that our programmers from Lita, who worked for 14$ an hour did a great job, And the engine will be up within the next couple'a months. if he did such a great job, why will it only be online in a couple of months? (sorry, couldn't resist). and I hope all these HiTech Bubble Programmers here who wont move themselfs for less then 50$ an hour for some things, i would charge you tripe that. others, i would do for free. what does that prove? are enjoying sitting at home on their fired/about to be fired ass and enjoying the view :) it's a free market. you are welcome to hire or not hire anyone you want. i do, however, discussing your hiring practices in such a manner on the list, shall we say, childish? I know its offtopic -- sorry about that! yeah, so am i. well, not really. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: Care and Feeding of Stupids (was: Re: stupid me)
On Sat, 29 Dec 2001, Nimrod Simba Carmi wrote: [major snippage. i'm no longer in the flaming mood] it's a free market. you are welcome to hire or not hire anyone you want. i do, however, discussing your hiring practices in such a manner on the list, shall we say, childish? I dont think so. I was responding to someone else talking about the way newbies are attacked when they come in here. I think this list has alot to only if they make a mess. work on if we're what's supposed to represent the israeli linux users. If say what? this list doesn't represent anyone and anything. at most, it represents its subscribers. we wanna ever get everyone to use linux, we have to be more patient to others and get rid of all this ideology coz it screws our point ! i'm not going there. open source vs free software discussions will be dealt with marc's scsi cable! -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
SourceForge Project Approved (fwd)
following the discussion we've had, i opened a sourceforge project for iglu. the project shall serve as: 1. a cvs repository for configuration files, scripts and any other data which should be both backed up and open for the public's view. 2. a TODO list. people with sourceforge ids who wish to be added as developers to the project should let me know, with their sf ids. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ -- Forwarded message -- Date: Wed, 26 Dec 2001 08:04:52 -0800 From: [EMAIL PROTECTED] To: [EMAIL PROTECTED] Subject: SourceForge Project Approved Your project registration for SourceForge has been approved. Project Full Name: The IGLU Server Project Unix Name: igludev CVS Server: cvs.igludev.sourceforge.net Shell/Web Server: igludev.sourceforge.net Your DNS will take up to a day to become active on our site. While waiting for your DNS to resolve, you may try shelling into shell.sourceforge.net and pointing CVS to cvs.sourceforge.net. If after 6 hours your shell/CVS accounts still do not work, please open a support ticket so that we may take a look at the problem. Please note that all shell/CVS accounts are closed to telnet and only work with SSH1. Your web site is accessible through your shell account. Please read site documentation (see link below) about intended usage, available services, and directory layout of the account. Please take some time to read the site documentation about project administration (http://sourceforge.net/docs/site/). If you visit your own project page in SourceForge while logged in, you will find additional menu functions to your left labeled 'Project Admin'. We highly suggest that you now visit SourceForge and create a public description for your project. This can be done by visiting your project page while logged in, and selecting 'Project Admin' from the menus on the left (or by visiting https://sourceforge.net/project/admin/?group_id=42759 after login). Your project will also not appear in the Trove Software Map (primary list of projects hosted on SourceForge which offers great flexibility in browsing and search) until you categorize it in the project administration screens. So that people can find your project, you should do this now. Visit your project while logged in, and select 'Project Admin' from the menus on the left. Enjoy the system, and please tell others about SourceForge. Let us know if there is anything we can do to help you. -- the SourceForge crew = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: Guy's Rants about volunteerism
On Sun, 23 Dec 2001, Omer Zak wrote: I suggest that the existing www.iglu.org.il volunteer maintainers make information available, to make it easier for other people to volunteer as follows: 1. Manage it like a Free Software project - make the various scripts, and documentation of the working of the Web site, available for read-only access via the WWW and/or FTP. 2. Build a TODO list of tasks which need to be performed, both regularly and one-time. Anyone who has a free moment may go over the TODO list and do something from it, without further commitment. i think this is an excellent idea. if no one objects, i'll set up a project on sourceforge (savannah?) for iglu's maintenance. all scripts, configuration files, mailing list and todo lists should go through there. Thus, anyone who has a free moment and an itch to scratch, can retrieve the relevant scripts, modify them and E-mail them to the webmaster (who will act as Linus - decide what patches to accept and what to reject). better have a mailing list, a-la iglu-web for these things, to keep everyone in the loop. someone should have the final deciding power, though. omerm? guy? - mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: biditext
On 23 Dec 2001, Erez Doron wrote: how do i get biditext to switch from visual to logical and vice-versa ? [i'm assuming you mean from rtl to ltr here?] that would depend on which version of biditext you're using. if you're using the old biditext, i really dont know. if you're using the latest experimental biditext, part of haifux' r2l project, you can use the set of 'r2l' tools which include a gnome panel applet and a command line utitlity to do it. the latest biditext is available at http://vipe.technion.ac.il/~mulix/r2l/biditext-0.9.6.tar.gz the latest r2llib, which it requires, is available at http://vipe.technion.ac.il/~mulix/r2l/r2llib-0.34.tar.gz r2llib includes a small utility called (how unsurprisingly) 'r2l', which is a command line utility to set the direction of the text (ltr to rtl) and the base direction. it also requires fribidi, which is available from sourceforge. tzafrir, where is the canonical distribution site for everything else? http://linuxclub.il.eu.org/R2L/index.html looks rather out of date. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: Guy's Rants about volunteerism
On 23 Dec 2001, Oleg Goldshmidt wrote: mulix [EMAIL PROTECTED] writes: better have a mailing list, a-la iglu-web for these things, to keep everyone in the loop. someone should have the final deciding power, though. omerm? guy? Why a-la? IGLU-web is there, just dormant, isn't it? Or does the sourceforge model mandate a sourceforge-related mailing list? i dont think there's any problem with using iglu-web as it is. sourceforge offer to setup mailing lists for you, but they do not mandate their use. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: Guy's Rants about volunteerism
On Sun, 23 Dec 2001, Ely Levy wrote: I think it's a great idea also but I think sourceforge project is really not it. we don't need any sf project just a little todo list running on iglu web site CVS. offsite backup. no more disappearing mirror scripts or mysteriously changed configuration files. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: C++: a gdb problem when debugging multiple files source program.
On Sat, 22 Dec 2001, Shaul Karl wrote: My overall impression so far is that the the ability to run substantial code before main get started is quite problematic. However I am as well it should be! you are aware, i hope, of the static initializer problem? http://www.codeguru.com/cpp/tic/tic0114.shtml currently programming only the initialization of my C++ program, which is why I only encounter difficulties with the initialization of a C++ program. but why does initialization of your code need to take place before main() begins?? note, if it doesn't, and i'm butting in at the middle of the thread, apologies. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: http://www.linux.org.il/ rants
On Sun, 23 Dec 2001, Shachar Shemesh wrote: Actually, I'm going over BugTraq daily anyways. What system is Iglu running anyway? RedHat? redhat 6.2. that would mean the initial update to bring it up to date will be rather massive, unless someone's been taking care of it in the meantime? -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
ANN: syscalltrack version 0.66 released
syscalltrack-0.66, the 4th _alpha_ release of the linux kernel system call
tracker, is available. syscalltrack supports both versions 2.2.x and 2.4.x
of the linux kernel. The current release contains some major enhancements,
and various bug fixes and code cleanups.
* What is syscalltrack?
syscalltrack is a linux kernel module and supporting user space
environment which allow interception, logging and possibly taking
action upon system calls that match user defined criteria
(syscalltrack can be thought of as a sophisticated, system wide
strace).
* Where can i get it?
Information on syscalltrack is available on the project's homepage:
http://syscalltrack.sourceforge.net, and in the project's file
release.
You can download the source directly from:
http://prdownloads.sourceforge.net/syscalltrack/syscalltrack-0.66.tar.gz
* Call for developers:
The syscalltrack project is looking for developers, both for kernel
space and user space. If you want to join in on the fun, get in touch
with us on the 'syscalltrack-hackers' mailing list
(http://lists.sourceforge.net/lists/listinfo/syscalltrack-hackers).
* License and NO Warrany
'syscalltrack' is Free Software, licensed under the GNU General Public
License (GPL) version 2. The 'sct_ctrl_lib' library is licensed under
the GNU Lesser General Public License (LGPL).
'syscalltrack' is in _alpha_ stages and comes with NO warranty.
If it breaks something, you get to keep all of the pieces.
You have been warned (TM).
Happy hacking and tracking!
===
Major new features for 0.66
---
* Support for tracking some socket calls (e.g. 'socket', 'listen', 'accept',
'connect') - yet still without the ability to match against the address
that a socket connects to.
* Support for 'after' rules (i.e. rules that are matched right after a syscall
is invoked, and thus can match and log also the syscall's return value).
This in addition to the existing 'before' rules (which are matched right
before entering into the system call). Defining if a rule is a 'before'
or 'after' rule is done using the 'when' keyword. Also, log formats can be
specified seperatly for the 'before' rules and for the 'after' rules.
Note: syntax for specifying a 'log_format' in the config file has changed.
please look at doc/sct_config_manual.txt for details.
* Support for an 'in' operator in filter expressions (e.g:
filter_expression { PARAMS[1] in (passwd, nsswitch.conf) }
With strings it looks for a substring match. With numbers it looks for
an exact match.
* Optimization - variables values are now calculated only when they are used
(using a callback mechanism) - rather then all values assigned before invoking
the rule matching engine.
* Modified the behaviour of unregistering system calls that are 'busy' -
they are fully unregistered by 'sct_rules.o', so it could be unloaded at will.
However, 'syscall_hijack.o' unregisters them without yet reducing its module
use count - it'll do that when the 'busy' syscall invocation(s) return.
Some syscalls may be blocking for days (e.g. sshd version 1.X blocks on
'accept' until a client connects to it, which could be days) - and not
allow unloading 'syscall_hijack.o' - but it won't incur any overhead on
new system call invocations.
major bug fixes for version 0.66:
* Quoted strings in filter expressions could not contain any special
characters (e.g. dot, equals sign and other operators, brackets, etc).
now they can, and they may also contain escaped double quotes, e.g.:
Tom said \hello there!\
* Fix for a potential reference-count breaking problem in syscall_hijack.
* Fixes for potential memory leaks in the rule engine and filters code.
* The 'tester' stability testing program now only prints real error messages,
so its output can actually be read.
* Various other minor bug fixes, as well as various code rewrites, aggregating
variables into structures, etc.
===
--
mulix
http://vipe.technion.ac.il/~mulix/
http://syscalltrack.sf.net/
=
To unsubscribe, send mail to [EMAIL PROTECTED] with
the word unsubscribe in the message body, e.g., run the command
echo unsubscribe | mail [EMAIL PROTECTED]
Re: cron question
On Sun, 16 Dec 2001, Ishai Parasol wrote: What can I do if I want to run a script on a regular basis using cron, but I need it between the system's crons (lets say every 3 hours) ? sorry, but what exactly do you mean? if you mean that you need it to run every three hours, it's very easy to do - man 5 crontab. if you mean that you need to run it as a user as well as having it run as a cron job, just put it someplace accessible to both cron and the user. if you need it to run at different times (every two to five hours from the last time it was run, depending on some parameter, for example), just do a neat hack and schedule it to run *once* at some start time, but in each invocation of the script, the script should schedule itself to run sometime in the future. this can barf if the script dies unexpectedly without having a chance to reschedule itself, so make sure to have it restarted once a day, if necessary. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: enabling ipchains/iptables
On Fri, 14 Dec 2001, TCL wrote: hello i have slack 8 with kernel 2.4.5 in the last week i got my modem working with linux and now is the time to set up a firewall ruleset actually, the correct time to do it was *before* you got your modem to work and connected for the first time. never underestimate the power of the cracker with a scanner. i have both ipchains and iptables installed, but unfurtunally, my kernel is compiled with no support to both is it possible to enable support without recompiling the kernel? nope. if not, how can i make sure all the current options compiled in the kernel will stay the same except for the firewall ones? i remember there was a way to make a config file of the current configuration, but i forgot it and, if i use that method, what line do i need to change? here's what i do. cd /path/to/old/kernel copy .config /someplace/safe cd /path/to/new/kernel cp /someplace/safe/.config ./ make oldconfig [this will only ask you questions which are new to this kernel version] make menuconfig [now change whatever needs to be changed] continue as usual. [dep, bzImage, etc]. note that this depends on having the .config of your current kernel. if it's a distro kernel, you might want to take the hour or two to reconfigure it, since distro kernels are built to suit the widest range of hardware they might encounter - almost everything is compiled, usuaully as modules. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: (OT) ISP's that don't allow relaying
On Fri, 14 Dec 2001 [EMAIL PROTECTED] wrote: SS too much :). You can try and get Actcom to do an SMTP after POP SS arrangment. In any case, don't get mad at them for not relaying. SS Relaying is BAAAD. Actually, relaying for own client's IP is good. That's the part of ISP service. And ISP can easily know for any given IP if it is their IP or not. And AFAIK most ISPs in Israel act exactly this way (I'm not sure particullary for Actcom, though). actcom does it as well, of course, but the case in point is when a client has an *email only* account, which means the client must be connected through some other isp, which means the ip is *not* one of actcom's ips - hence, no relaying. -- mulix http://vipe.technion.ac.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: [Slightly Off Topic] processor speed
On Wed, 5 Dec 2001, Dan Kenigsberg wrote: I admit this is slightly Meaningless Numbers Calculus, but I have to measure a certain computational task, and come up with a number per architecture. I could use total process time, but dividing it with the processor speed to acquire a measure of total cpu cycles seems more appropriate. as for process time, you can use getrusage() to get the total user time and system time. Returning to my problem, I can use /proc/cpuinfo (in fact, I am). However, on Sun/Alpha the best I can do is run psrinfo, but it is problematic since those systems are in some cases assymetric - and I cannot tell on which cpu my proccess is running. use inline asm to get cpu speed? i would think that all cpus have an instruction to get this information. you can look at the kernel's arch/ subdir to see how it's calculated for the different architectures. note that this is most likely a can of worms, unless you only need to support a limited subset of processors architectures. -- mulix http://www.pointer.co.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: An Obscure Bug
On Tue, 4 Dec 2001, Shlomi Fish wrote: I discovered an obscure bug which I am able to reproduce only under very rare circumctances. Here are the instructions for re-producing it: Download http://vipe.technion.ac.il/~shlomif/obscure-bug.tar.gz Unpack and type make. Then type source test.sh at the command line and press enter. Voila - a segfault. Now, ./test.sh doesn't cause it. this is imperative - 'source test.sh' causes it, but './test.sh' doesn't? The program in question was written by me (it's version 1.11.24 of Freecell Solver), and is compiled with the -O3 flag by gcc. If it's compiled with -O2 it doesn't segfault. Nor does it when compiled with kgcc. Moreover, after I made a very small change in the program, it did not segfault either. Finally, running it under gdb cause it not to crash either. I'm using Mandrake 8.1 with KDE 2.2.1 and all of the up-to-current updates. My kernel is 2.4.13-12mdk from cooker. which gcc? it ran fine with gcc2.91.66, gcc2.95.2 and gcc3.0 (using 'source test.sh'). . -- mulix http://www.pointer.co.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: DON'T USE 2.4.15(!!) (was: Re: Local SuSE mirror)
On Tue, 4 Dec 2001, guy keren wrote: On Mon, 3 Dec 2001, Gold Edward wrote: We ended up on RH7.1 (2.4.10 updated afterwards to 2.4.15). 2.4.15 has a bug that may (and does) cause it to trash filesystems. don't use this version - immediastly reset your system, boot with the older kernel, and upgrade to some other version (2.4.13, or if 2.4.16 is out - then probably 2.4.16. 2.4.14 has a bug that does not allow using loopback mounts, unless you add a few lines of fixed code). 2.4.16 is out and is stable on all accounts. 2.4.17pre2 is out as well, if you want to live on the (stable) bleeding edge. as for 2.4.15, dont forget to force fsck after you boot from it to a new kernel. otherwise you could get *silent* data corruption - not a nice thing by any stretch of the imagination. for more information on this bug, peruse this list's archives, the linux kernel mailing list archives, slashdot or any other linux news source of your liking. p.s. to fix 2.4.14 you have to remove two lines, not add. i know, i konw, nit picking. can't help it. -- mulix http://www.pointer.co.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: Cox kernel?
On Sat, 1 Dec 2001, Hetz Ben-Hamo wrote: The 2.4.x kernel development has been moved from Alan to Marcello. Linus will only announce new kernel 2.4.x releases, but won't touch any code actually... ? 2.4.16 was released by marcello. so was 2.4.17pre2. linux didnt have anything to do with these, publicly. So Rik's VM implementation is out, and other Linus stuff that were in the kernel (while they weren't in Alan's ac-tree) has been removed in favour of.. you guessed it - Alan's tree.. ??? rik's vm is indeed out (although he is maintaining it himself), but marcello started working from 2.4.15, which was linus' tree, not alan's tree. i think you are mistaken here. to answer the original question, the important bits in -ac have been merged into -linus before 2.4.15. as to the rest, i dont know. maybe it's 2.5.x material. is there anything specific you're interested in, tzahi? -- mulix http://www.pointer.co.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: Linux compatible modem
i have added this answer to iglu's FAQ, at http://www.iglu.org.il/faq/cache/170.html, under 'networking'. (tzafrir, please change the title from 'New Item' to something saner. it looks like only you have permissions to do so?) On Fri, 30 Nov 2001, Gal Goldschmidt wrote: Hi, I did a major research about this topic: Hardware: External 3Com(USR) Modem Serial or USB you might be able to get some dealer to order it for you. They also have a hardware PCI, but no one bring it to Israel ( as far as I know). Zoom USB and Zoom PCI hardware, both can be ordered from www.abnet.co.il. It will take them some time since it's not on the shelf. IBM PCI Hardware modem 33L4618 ( The one I finally ordered) can be ordered from any IBM dealer, works great! If you want something tomorrow and not 3 weeks from order time: Software: Intel modem, from www.ksp.co.il the PCI not the nice USB has a driver http://developer.intel.com/design/modems/products/md563x.htm Not tested by me yet, but I plan to test it. Bye Gal -- mulix http://www.pointer.co.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: checking the functioning of an ipchains module
On Sat, 1 Dec 2001, guy keren wrote: On 30 Nov 2001, Noam Meltzer wrote: I guess you didn't really understand what i wanted. I don't want to see that the module is loaded. I want to see what is it doing while it's running. what its doing has different interpretations. if it is 'understanding how it works' - use the source, luke. if its seeing which packets get NATed - i _think_ there's an option to enable some kind of debug code in netfilter's code which _could_ help. or its something else? you might run a sniffer before the NAT box and after the NAT box, look at the output, and begin analising it ;) be carefull... there be dragons here (in relation to the analysis part). there is *supposed* to a file in /proc, which tells you which connections are being nat'ed on your box, /proc/net/ip_masquerade. for some reason, it's not there on my linux router. any ideas where it's gone? also (2 questions for the price of one email), i'm looking to implement traffic limiting on the linux router for internal users (bofh? me? never. what was your user name again?). what tools am i looking for? kernel 2.4.16, approximately latest iptables. -- mulix http://www.pointer.co.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: checking the functioning of an ipchains module
On Sat, 1 Dec 2001, Oded Arbel wrote:
I don't have that file, but I have /proc/net/ip_conntrack which under
correct analyzis will yield the list of NATed connections.
(kernel 2.4.13, iptables)
i must have looked at it the other time when no internal client was
connected, since i only saw the linux router's ip in there and assumed
it was only for local connections.
anyway, thanks. here's a small script i wrote now to only show you
tcp connections where the src or dst match a certain regexp [1]
[1] yes, i know grep can do it too. dont you think i would've used it,
if it suited my purpose? the owls are calling again, and the script is
not what it seems.
#!/usr/bin/perl -w
#
# $Id: listcons.pl,v 1.1 2001/12/01 12:01:37 mulix Exp $
#
# print all tcp connections going through the box. if a parameter
# is given, only print a connection where the src or dst is this regexp.
# mulix [EMAIL PROTECTED]
#
# fields explanation at
# http://lists.samba.org/pipermail/netfilter/2001-February/007830.html
#
use strict;
my $proc_file_name = /proc/net/ip_conntrack;
my @connections;
my $filter = $ARGV[0];
open (PROC, $proc_file_name) or die couldn't open $proc_file_name - $!;
while (PROC) {
if (/^tcp/) { #only handle tcp connections for now
if (/^\s*(\S*)\s*(\d*) (\d*) (\S*) src=([\d\.]*) dst=([\d\.]*) sport=(\d*)
dport=(\d*) src=([\d\.]*) dst=([\d\.]*) sport=(\d*) dport=(\d*)/) {
my $con_stat = {
PROTO = $1,
PROTO_NUM = $2,
TTL = $3,
TCP_STATUS = $4,
SRC1 = $5,
DST1 = $6,
SPORT1 = $7,
DPORT1 = $8,
SRC2 = $9,
DST2 = $10,
SPORT2 = $11,
DPORT2 = $12,
};
push @connections, $con_stat;
} else {
print parsed unknown line: $_\n;
}
}
}
print_connections();
sub print_connections()
{
my $c;
foreach $c (@connections){
if (defined $filter) {
next unless (($c-{SRC1} =~ /$filter/) or ($c-{DST1} =~ /$filter/));
}
print $c-{PROTO}: $c-{SRC1}:$c-{SPORT1} = ,
$c-{DST1}:$c-{DPORT1}\n;
}
}
--
mulix
http://www.pointer.co.il/~mulix/
http://syscalltrack.sf.net/
=
To unsubscribe, send mail to [EMAIL PROTECTED] with
the word unsubscribe in the message body, e.g., run the command
echo unsubscribe | mail [EMAIL PROTECTED]
RE: Cox kernel?
On Sat, 1 Dec 2001, Tzahi Fadida wrote: [ my interpretation: which kernel should i use for maximum stablity, as a router? ] 2.4.16 seems fine, but it hasn't been out long enough to be sure. my advice to you is to use the latest distro kernel, because you can be reasonably sure that kernel version had at least some QA done to it. if you're running redhat, that would be 2.4.9-soemthing. as to how well netfilter will play with it - i have no idea. = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: Cox kernel?
On Sat, 1 Dec 2001, Hetz Ben-Hamo wrote: So Rik's VM implementation is out, and other Linus stuff that were in the kernel (while they weren't in Alan's ac-tree) has been removed in favour of.. you guessed it - Alan's tree.. ??? rik's vm is indeed out (although he is maintaining it himself), but marcello started working from 2.4.15, which was linus' tree, not alan's tree. i think you are mistaken here. Umm, sorry, you might want to look real close into the the sources of 2.4.14 and 2.4.15 - all Alan's changes from Linus tree are in the 2.4.15 tree (look at the sound drivers for example - You could once make a driver with the kernel to play directly at 44Khz - now you can only 48Khz and downmix to 44Khz with a user space app) guess you didnt read all of what i wrote. here it is: to answer the original question, the important bits in -ac have been merged into -linus before 2.4.15. as to the rest, i dont know. maybe it's 2.5.x material. is there anything specific you're interested in, tzahi? obviously it's impossible for all of -ac to have been merged, since it also included rik's vm... -- mulix http://www.pointer.co.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: checking the functioning of an ipchains module
On Sat, 1 Dec 2001, Tzafrir Cohen wrote: On Sat, 1 Dec 2001, mulix wrote: also (2 questions for the price of one email), i'm looking to implement traffic limiting on the linux router for internal users (bofh? me? never. what was your user name again?). what tools am i looking for? Have you looked at the advanced routing howto? shame on me for asking before doing some research, but it really was a spur of the moment question. i'll check out the advanced routing howto, but in the mean time, a quick google search found this article: http://linux.oreillynet.com/pub/a/linux/2000/08/24/LinuxAdmin.html, which points to the IP-QoS howto: http://qos.ittc.ukans.edu/howto/. the article is short but has a few good pointers. the howto i'm reading now. -- mulix http://www.pointer.co.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: adsl keep alive script
On Thu, 29 Nov 2001 [EMAIL PROTECTED] wrote: Hello all, I'm trying to write a script which will keep my adsl connection up all the time. good for you, although many such scripts have already been written. it's a useful excersize, if nothing else. I tried something like running : while 1 rm -rf /var/run/pp* pptp parameters nodetach end but from some reason, whenever i ran it, it worked for about 10 minutes, and after ten minutes, the pptp was detached and alot of pptp process started ( even that the connection was still alive ). that pptp line looks oxymoronic. how are 'nodetach' and '' supposed to work together? i use a script written by haif gelfenbeym, called adsl-connect. see it here: http://www.mail-archive.com/linux-il%40cs.huji.ac.il/msg12428.html etzion also wrote a good script, but i dont have a url handy. etzion? studiously avoiding all political discussions, mulix -- mulix http://www.pointer.co.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: adsl keep alive script
On Fri, 30 Nov 2001, Ghiora Drori wrote: I probably could use SIGTERM but I like using an axe... that's all nice and well, if you're an axe murderer. for killing programs, it's *really* prefered to first try to gently smother them with a pillow (kill) and only then resort to somewhat more violent means (kill -9). why? because pretty much any non trivial, well coded program needs to perform cleanup on shutdown. kill -9 will won't let that happen, possibly causing problems the next time the program comes up or even before (think system wide resources which are not cleaned up on process termination, which the poor program might leave behind because you splattered its brains all over the wall with a well timed axe stroke, err, kill -9). -- mulix http://www.pointer.co.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: fork. system login failure
On Mon, 26 Nov 2001, Stiven Andre wrote: I have some strange probelm. I have server running rh7.2 that works 24hours a day and after adding some cron jobs I always recieve the following: When the server runs about 10-20 hours I recieve an error sh: fork Resource temporary unavilable for the first 3-4 commands I enter. after that it goes normaly. But if I will leave server alone for 40-50 hours I can't even login (Using the console not remotly). When I type username and pass it logs me in for 2-3 secs after that writes something (not enough time too see what) and logs me out. The server still functionaling normaly all the services wotk as it should but I can't login. sounds to me like your cron jobs are creating processes which are then left hanging around, causing later forks to fail. the is a limit on the number of processes a system can have. what do your cron jobs do? -- mulix http://www.pointer.co.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: C program segfaults when run, not under GDB
the problem here isn't a buffer overflow, i'm afraid. the problem here
is that s1 and s2 are in a read only section of the executable, and
trying to modify s1 is forbidden.
consider this snippet, which produces a SIGSEGV as well:
void foo(char s[]){
s[0] = 'b';
}
int main()
{
foo(a);
return 0;
}
On Sun, 25 Nov 2001, Max Kovgan wrote:
this line:
if (factor!=0) s1[c1]=s1[c1+factor];
tries to see behind the s1's end after the index (length_of_s1 - factor)
is reached.
this causes SIGSEG
it's not nice to ask people to debug your own homework :))
bye
-=O0~~O0=-
Beware the Jabberwock, my son!
The jaws that bite, the claws that catch!
Beware the Jubjub bird, and shun
The frumious Bandersnatch!
[L.Carrol Jabberwacky]
On Sun, 25 Nov 2001, Eugene Romm wrote:
Hello.
I've written a procedure that's supposed to remove all occurances of
string2 from string1 (parameters).
For reasons I do not understand, the program compiles but segfaults when
run from the command prompt, but silently executes without a warning
when run under GDB. Attached is the program. Segfault occurs on line 29,
as far as I can tell.
--
mulix
http://www.pointer.co.il/~mulix/
http://syscalltrack.sf.net/
=
To unsubscribe, send mail to [EMAIL PROTECTED] with
the word unsubscribe in the message body, e.g., run the command
echo unsubscribe | mail [EMAIL PROTECTED]
Re: Alcatel ADSL modem, Linux and Bezeq ADSL
On Tue, 20 Nov 2001, Sagi Bashari wrote: Hi Omer On Tuesday 20 November 2001 14:53, Omer Zak wrote: The MTU is set correctly (1500 in Ethernet interfaces, 1452 in pptp/ppp0). Did you reduce the MTU to 1452 on the internal hosts too? the adsl howto explains how to do it in linux and windows. Can you surf to the problematic sites through the gateway? if you can't, you might want to check if ecn is enabled on the linux router. some (misconfigured) firewalls block ecn. check /proc/sys/net/ipv4/tcp_ecn. if it exists, try echo 0 tcp_ecn. btw, the --quirks patch has been accpeted to cvs pptp last night... -- mulix http://www.pointer.co.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: C++: Problem with overloading a constructor when splitting a
On Sun, 18 Nov 2001, Dan Kenigsberg wrote: This may seem an ugly feature of C++, but in fact it is better than the C counterpart - macros. In fact, writing the implementation of inline functions in header file is a beautiful gem, comparing to writing the implemetation of calss templates in header files. Yuck. actually, writing the implementation of class templates in header files is *not* a language feature, it's a limitation of current (?) compiler technology, which can not handle templates being split up in different compilation units. [ i would supply more details, but couldnt find any references on google. if anyone can shed more light, please do. ] -- mulix http://www.pointer.co.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
sourceforge mysteries
hello, if anyone know the answer to these mystery of the universe questions, i will be most obliged. all deal with sourceforge's services for project administrators. i can say a lot of good things about sourceforge, but 'documentation' and 'user interface' will NOT be mentioned there. 1. how do you (the project adminstrator) control who has write access to your cvs repository? 2. how do you remove developers from your project? 3. how do you remove a sourceforge user completely? thanks in advance. -- mulix http://www.pointer.co.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: sourceforge mysteries
after bad mouthing sourceforge earlier today, i now come to praise them for their quick reponse to my support requests. On Sun, 18 Nov 2001, Ely Levy wrote: On Sun, 18 Nov 2001, mulix wrote: hello, if anyone know the answer to these mystery of the universe questions, i will be most obliged. all deal with sourceforge's services for project administrators. i can say a lot of good things about sourceforge, but 'documentation' and 'user interface' will NOT be mentioned there. actually I found their docs quite helpfull 1. how do you (the project adminstrator) control who has write access to your cvs repository? Developers on your project are automatically granted write access to your project CVS repository via SSH. Anonymous, read-only pserver-based access to your repository is provided to the general public true. more information available at: http://sourceforge.net/tracker/?func=detailatid=21aid=483030group_id=1 2. how do you remove developers from your project? Removing developers from your project Click Remove button near the developer's name on Project Admin page (yes, that button looks as a trashcan). Project administrators cannot be removed in this manner. To remove an administrator, you should reset his administrator privilege first. You cannot remove this privilege for yourself, and should ask the other admin to perform the removal operation. If you're the only admin, you may: doh, i missed this one. more information at: http://sourceforge.net/tracker/?func=detailatid=21aid=483032group_id=1 3. how do you remove a sourceforge user completely? this one was harder to find in the docs if you ever find out..:) thanks to sf support, i did. you log in as that user and send a support request to be removed. more info here: http://sourceforge.net/tracker/?func=detailatid=21aid=483032group_id=1 -- mulix http://www.pointer.co.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: sourceforge mysteries
On Sun, 18 Nov 2001, Shaul Karl wrote: hello, if anyone know the answer to these mystery of the universe questions, i will be most obliged. all deal with sourceforge's services for project administrators. i can say a lot of good things about sourceforge, but 'documentation' and 'user interface' will NOT be mentioned there. Although it does not answers your questions (at least not directly), you might be interested in http://mailman.fsfeurope.org/pipermail/announ ce/2001-November/28.html. I wonder what do you think about it. i've read it. also the big thread on advogato which probably sparked this announcement, at http://www.advogato.org/article/376.html my opinion has always been rather pragmatic on such issues. any new project i set up now will be on savannah or on one of my of my own servers, because in principle i agree with sourceforge's detractors. syscalltrack, however, will remain on sourceforge until it will no longer be convenient to have it there, or until me, or one of the other admins, finds the time to move it elsewhere. -- mulix http://www.pointer.co.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: loop.o problem
On Wed, 14 Nov 2001, Hetz Ben Hamo wrote: Hi people, I'm trying to compile loop.o module inside my kernel, and I get this weird message - does someone knows about this anything? hetz]# modprobe loop /lib/modules/2.4.14/kernel/drivers/block/loop.o: unresolved symbol deactivate_page /lib/modules/2.4.14/kernel/drivers/block/loop.o: insmod /lib/modules/2.4.14/kernel/drivers/block/loop.o failed /lib/modules/2.4.14/kernel/drivers/block/loop.o: insmod loop failed I checked all the usual places - didn't find anything special you didnt check well enough. there were at least 50 messages on linux kernel in regards to this problem. short answer: get 2.4.15pre-something (some of them were expermintal, so be careful which pre you grab) real answer: --- linux-2.4.14-broken/drivers/block/loop.cThu Oct 25 13:58:34 2001 +++ linux-2.4.14/drivers/block/loop.c Mon Nov 5 17:06:08 2001 @@ -207,7 +207,6 @@ index++; pos += size; UnlockPage(page); - deactivate_page(page); page_cache_release(page); } return 0; @@ -218,7 +217,6 @@ kunmap(page); unlock: UnlockPage(page); - deactivate_page(page); page_cache_release(page); fail: return -1; -- mulix http://www.pointer.co.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: [pptp-devel] RFC: required patch to pptp 1.0.3 for israeli adsl service
attached is a revised patch for bezeq support, relative to current pptp
cvs. bezeq support is now a run time mechanism, via a
--quirks=BEZEQ_ISRAEL switch. for this, i implemented a general quirks
mechanism, which allows overriding the out_call_rqst packet
construction, start_ctrl_conn packet construction and the set_link_info
packet construction.
comments are welcome, of course.
* the quirks mechanism is general enough to be useful to other people,
but not too general to be cumbersome to use. i added hooks for the exact
places i need them, other hooks can be added later, if needed.
* i'm not too happy with the interface to the quirks mechanism, nor with
hardware/isp distinction. perhaphs there should be --hardware-quirks
and --isp-quirks? for now, i index the quirks table based on isp (bezeq)
only, since the fix has been shown to work even with non quirky
hardware.
* this is only a rough patch. It Works For Me(tm).
diff -ur --new-file pptp-linux/Makefile pptp-mulix/Makefile
--- pptp-linux/Makefile Fri May 11 10:48:14 2001
+++ pptp-mulix/Makefile Wed Nov 14 12:12:40 2001
@@ -17,12 +17,18 @@
PPTP_BIN = pptp
PPTP_OBJS = pptp.o pptp_gre.o ppp_fcs.o \
pptp_ctrl.o dirutil.o vector.o \
-inststr.o util.o version.o
-PPTP_DEPS = pptp_callmgr.h pptp_gre.h ppp_fcs.h util.h
+inststr.o util.o version.o \
+ pptp_quirks.o orckit_quirks.o
+
+PPTP_DEPS = pptp_callmgr.h pptp_gre.h ppp_fcs.h util.h \
+ pptp_quirks.h orckit_quirks.h
CALLMGR_BIN = pptp_callmgr
-CALLMGR_OBJS = pptp_callmgr.o pptp_ctrl.o dirutil.o util.o vector.o version.o
-CALLMGR_DEPS = pptp_callmgr.h pptp_ctrl.h dirutil.h pptp_msg.h vector.h
+CALLMGR_OBJS = pptp_callmgr.o pptp_ctrl.o dirutil.o util.o \
+ vector.o version.o pptp_quirks.o orckit_quirks.o
+CALLMGR_DEPS = pptp_callmgr.h pptp_ctrl.h dirutil.h pptp_msg.h vector.h \
+ pptp_quirks.h orckit_quirks.h
+
all: $(PPTP_BIN) $(CALLMGR_BIN)
diff -ur --new-file pptp-linux/orckit_quirks.c pptp-mulix/orckit_quirks.c
--- pptp-linux/orckit_quirks.c Thu Jan 1 02:00:00 1970
+++ pptp-mulix/orckit_quirks.c Wed Nov 14 13:21:14 2001
@@ -0,0 +1,86 @@
+/* orckit_quirks.c .. fix quirks in orckit adsl modems
+ *mulix [EMAIL PROTECTED]
+ *
+ * $Id$
+ */
+
+#include string.h
+#include netinet/in.h
+#include pptp_msg.h
+#include pptp_options.h
+#include pptp_ctrl.h
+#include util.h
+
+
+
+/* return 0 on success, non zero otherwise */
+int
+orckit_atur3_build_hook(struct pptp_out_call_rqst* packet)
+{
+
+unsigned int length = 10;
+
+struct pptp_out_call_rqst fixed_packet = {
+ PPTP_HEADER_CTRL(PPTP_OUT_CALL_RQST),
+ 0, /* hton16(call-callid) */
+ 0, /* hton16(call-sernum) */
+ hton32(PPTP_BPS_MIN), hton32(PPTP_BPS_MAX),
+ hton32(PPTP_BEARER_DIGITAL), hton32(PPTP_FRAME_ANY),
+ hton16(PPTP_WINDOW), 0, hton16(length), 0,
+ {'R','E','L','A','Y','_','P','P','P','1',0}, {0}
+};
+
+if (!packet)
+ return -1;
+
+memcpy(packet, fixed_packet, sizeof(*packet));
+
+log(%s called\n, __FUNCTION__);
+return 0;
+}
+
+/* return 0 on success, non zero otherwise */
+int
+orckit_atur3_set_link_hook(struct pptp_set_link_info* packet,
+ int peer_call_id)
+{
+struct pptp_set_link_info fixed_packet = {
+ PPTP_HEADER_CTRL(PPTP_SET_LINK_INFO),
+ hton16(peer_call_id),
+ 0,
+ 0x,
+ 0x};
+
+if (!packet)
+ return -1;
+
+memcpy(packet, fixed_packet, sizeof(*packet));
+return 0;
+}
+
+int
+orckit_atur3_start_ctrl_conn(struct pptp_start_ctrl_conn* packet)
+{
+struct pptp_start_ctrl_conn fixed_packet = {
+ {0}, /* we'll set the header later */
+ hton16(PPTP_VERSION), 0, 0,
+ hton32(PPTP_FRAME_ASYNC), hton32(PPTP_BEARER_ANALOG),
+ hton16(0) /* max channels */,
+ hton16(0x6021),
+ {'R','E','L','A','Y','_','P','P','P','1',0}, /* hostname */
+ {'M','S',' ','W','i','n',' ','N','T',0} /* vendor */
+};
+
+if (!packet)
+ return -1;
+
+/* grab the header from the original packet, since we dont
+ know if this is a request or a reply */
+memcpy(fixed_packet.header, packet-header, sizeof(struct pptp_header));
+
+/* and now overwrite the full packet, effectively preserving the header */
+memcpy(packet, fixed_packet, sizeof(*packet));
+return 0;
+}
+
+
diff -ur --new-file pptp-linux/orckit_quirks.h pptp-mulix/orckit_quirks.h
--- pptp-linux/orckit_quirks.h Thu Jan 1 02:00:00 1970
+++ pptp-mulix/orckit_quirks.h Wed Nov 14 12:34:21 2001
@@ -0,0 +1,27 @@
+/* orckit_quirks.h .. fix quirks in orckit adsl modems
+ *mulix [EMAIL PROTECTED]
+ *
+ * $Id$
+ */
+
+#ifndef INC_ORCKIT_QUIRKS_H_
+#define INC_ORCKIT_QUIRKS_H_
+
+#include pptp_options.h
+#include pptp_ctrl.h
+#include pptp_msg.h
+
+/* return 0 on success, non zero otherwise */
+int
+orckit_atur3_build_hook(struct
RFT: pptp patch for bezeq against current cvs pptp
hello, linuxers, if anyone wants to help test the new pptp patch against current cvs pptp, you can grab it from http://www.pointer.co.il/~mulix/. there's both a diff and a .tar.gz. this patch has run time support for bezeq, instead of compile time support like we used to have. it also appears to connect much faster, due to improved code in the cvs pptp version. you run it like this: /pptp 10.0.0.138 --quirks=BEZEQ_ISRAEL bla bla bla if you test it, please let me know if it works for you. if you're feeling extra bold, let me know if it works with and without the --quirks option. it works for me and for several other people, and it's fundamentally the same code as the pptp-1.0.3 patch, so it should be safe enough. then again, it should be safe enough are famous last words. thanks, -- mulix http://www.pointer.co.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: [pptp-devel] RFC: required patch to pptp 1.0.3 for israeli adsl service
On Tue, 13 Nov 2001, James Cameron wrote: G'day mulix, good day, james. Thank you for your patch to pptp that adds support for Bezeq. thank you for maintaining pptp and thanks to the original author, for writing it. - I'm wondering if the #define should be something other than Bezeq. Have you heard if any other implementations need this change? Have you any idea what software or hardware Bezeq is using to deliver ADSL? no idea, but i think dani answered this point. - can this code change be added as a configuration option rather than an #ifdef? probably... i'll grab the latest pptp cvs source and make a patch later today. - changes are made to pptp_options.h for Bezeq, but some of them are ignored by the struct pptp_out_call_rqst packet changes. Would it be possible to define additional constants so that the code that initialises the struct has no #ifdef? In other words, can the changes be isolated somehow in order to simplify the code. would generating the packet at run time, based on the user supplied command line options, be acceptable? something like struct pptp_out_call_rqst packet; build_call_rqst_packet(packet); where build_call_rqst_packet would build the correct packet, based on whether the user supplied --bezeq or not. this deviates from the common style in the code for packet construction, which is the reason i didnt do it in the first place. - do the changes to add pptp_set_link need to be #ifdef'd? Has this been tested with conventional PPTP environments? since the only pptp environment i have is bezeq's adsl service, it hasnt. i wrote this specific bit of the patch, based on packet dumps of windows boxes. i remember set_link is defined in the rfc, but i dont think it's required. it's been a while since i read the rfc, though. - the changes to the error reporting are good, they improve the situation. good, that was the first thing we did ;) Patch relative to current CVS is attached. No significant changes made, just made sure it applies cleanly. thanks. -- mulix http://www.pointer.co.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: QT3 Installation
On Tue, 13 Nov 2001, Yoav Bornstein wrote: I am afraid there are no more details to give. It looks like everything is fine until it reachs qrichtext.cpp, then it's just looks like it stuck in an endless loop or something (believe me, the computer worked on it all night). I used the offical release which I've downloaded from ftp.trolltech.com. I'll give some more details about my computer and distro : IBM ThinkPad (Cel550,64RAM,...) Mandrake 8.1 compiler version? this is very important, actually. gcc -v should give you the details. -- mulix http://www.pointer.co.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
Re: experimental pptp-mulix rpms
On Mon, 12 Nov 2001, Yotam Rubin wrote: Just a question, is there any reason why the modified pptp can't be merged with the upstream pptp? It'll be a whole a lot cleaner. pptp-1.0.3 has a lot of the bugfixes that went into our 1.0.2 patch. the rest are (afaicr) small protocol modifications required for it to work here, minor debugging enhancements and country specific information. merging them with the upstream pptp will require either extensive pptp hacking, to provide a framework such changes can be made in, or some #ifdef ugliness. the first is possible, the second unlikely to be accepted (i know i wouldn't accept it if i was the pptp maintainer). what's required is for someone (probably me, i suppose) to send our patch to the pptp maintainers and ask for comments, to see what will be required for inclusion. i should have done it a long time ago, but it never got to the top of my TODO queue. i'll give it a temporary priority boost and do it later tonight. -- mulix http://www.pointer.co.il/~mulix/ http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
RFC: required patch to pptp 1.0.3 for israeli adsl service
hello,
i'd like to solicite comments for the attached patch to pptp-1.0.3. this
patch is required for pptp to work with the adsl equipment used by the
single adsl provider in israel. it's been in use for the last year and
works with all adsl modem types, while vanilla pptp-1.0.3 fails with at
least some modems.
i'm not suggesting for the patch to be applied in its current form, but
i would like to solicit comments on how it can be changed so that it
will be applied. in particular, how should the country specific
information be handled?
this patch is based on an earlier patch we did to pptp-1.0.2. both are
available at http://www.pointer.co.il/~mulix/. the original patch
(against 1.0.2) was derived from watching packet dumps of succesful
windows connections vs. failed linux connections, and the 1.0.3 patch
was a merge of the 1.0.2 patch with the original pptp 1.0.3.
diff -ur pptp-linux-1.0.3/Makefile pptp-mulix-1.0.3/Makefile
--- pptp-linux-1.0.3/Makefile Mon May 7 06:19:34 2001
+++ pptp-mulix-1.0.3/Makefile Fri Jun 8 23:41:01 2001
@@ -1,10 +1,11 @@
VERSION = 1.0.3
VERSION_DEFINE = '-DPPTP_LINUX_VERSION=${VERSION}'
+COUNTRY_DEFINE = '-DPPTP_BEZEQ_ISRAEL=1'
CC = gcc -Wall
DEBUG = -g
INCLUDE =
-CFLAGS = -O1 $(VERSION_DEFINE) $(DEBUG) $(INCLUDE) -DPROGRAM_NAME=\pptp\
+CFLAGS = -O1 $(VERSION_DEFINE) $(DEBUG) $(COUNTRY_DEFINE) $(INCLUDE)
+-DPROGRAM_NAME=\pptp\
LIBS =
LDFLAGS= -lutil
Only in pptp-mulix-1.0.3/: pptp
diff -ur pptp-linux-1.0.3/pptp.c pptp-mulix-1.0.3/pptp.c
--- pptp-linux-1.0.3/pptp.c Mon Apr 30 06:42:36 2001
+++ pptp-mulix-1.0.3/pptp.c Fri Jun 8 23:42:09 2001
@@ -42,6 +42,7 @@
void usage(char *progname) {
fprintf(stderr,
%s\n
+ patched by mulix [EMAIL PROTECTED] for Israel\n
Usage:\n
%s hostname [[--phone phone number] -- ][ pppd options]\n,
version, progname);
Only in pptp-mulix-1.0.3/: pptp_callmgr
diff -ur pptp-linux-1.0.3/pptp_ctrl.c pptp-mulix-1.0.3/pptp_ctrl.c
--- pptp-linux-1.0.3/pptp_ctrl.cMon Apr 30 06:42:36 2001
+++ pptp-mulix-1.0.3/pptp_ctrl.cFri Jun 8 23:55:01 2001
@@ -205,13 +205,25 @@
call-closure = NULL;
/* Send off the call request */
{
-struct pptp_out_call_rqst packet = {
+struct pptp_out_call_rqst packet =
+#ifndef PPTP_BEZEQ_ISRAEL
+{
PPTP_HEADER_CTRL(PPTP_OUT_CALL_RQST),
hton16(call-call_id), hton16(call-sernum),
hton32(PPTP_BPS_MIN), hton32(PPTP_BPS_MAX),
hton32(PPTP_BEARER_CAP), hton32(PPTP_FRAME_CAP),
hton16(PPTP_WINDOW), 0, 0, 0, {0}, {0}
};
+#else /* PPTP_BEZEQ_ISRAEL is defined */
+{
+ PPTP_HEADER_CTRL(PPTP_OUT_CALL_RQST),
+ 0, /* hton16(call-callid) */
+ 0, /* hton16(call-sernum) */
+ hton32(PPTP_BPS_MIN), hton32(PPTP_BPS_MAX),
+ hton32(PPTP_BEARER_DIGITAL), hton32(PPTP_FRAME_ANY),
+ hton16(PPTP_WINDOW), 0, hton16(PPTP_HOSTNAME_LEN),0, PPTP_HOSTNAME, {0}
+};
+#endif
/* fill in the phone number if it was specified */
if( phonenr ){
@@ -473,6 +485,22 @@
}
}
+static void
+pptp_set_link(PPTP_CONN* conn, int peer_call_id)
+{
+ struct pptp_set_link_info packet = {
+ PPTP_HEADER_CTRL(PPTP_SET_LINK_INFO),
+ hton16(peer_call_id),
+ 0,
+ 0x,
+ 0x};
+
+ if (pptp_send_ctrl_packet(conn, packet, sizeof(packet))) {
+ log(pptp_set_link() packet sending succesfull);
+ pptp_reset_timer();
+ }
+}
+
void pptp_dispatch_ctrl_packet(PPTP_CONN * conn, void * buffer, size_t size) {
struct pptp_header *header = (struct pptp_header *)buffer;
u_int8_t close_reason = PPTP_STOP_NONE;
@@ -639,8 +667,21 @@
if (call-state.pns == PNS_WAIT_REPLY) {
/* check for errors */
if (packet-result_code!=1) {
- /* An error. Log it. */
- log(Error opening call. [callid %d], (int) callid);
+ /* An error. Log it verbosely. */
+ unsigned int legal_error_value =
+ sizeof(pptp_general_errors)/sizeof(pptp_general_errors[0]);
+ int err = packet-error_code;
+ log(Error '%d' opening call. [callid %d],
+ packet-result_code, (int) callid);
+ log(Error code is '%d', Cause code is '%d', err,
+ packet-cause_code);
+ if ((err 0) (err legal_error_value)){
+log(Error is '%s', Error message: '%s',
+pptp_general_errors[err].name,
+pptp_general_errors[err].desc);
+ }
+
+
call-state.pns = PNS_IDLE;
if (call-callback!=NULL) call-callback(conn, call, CALL_OPEN_FAIL);
pptp_call_destroy(conn, call);
@@ -650,6 +691,7 @@
call-peer_call_id = ntoh16(packet-call_id);
call-speed= ntoh32(packet-speed);
pptp_reset_timer();
+ pptp_set_link(conn, call-peer_call_id);
if (call-callback!=NULL) call-callback(conn, call, CALL_OPEN_DONE);
log(Outgoing call established
Re: Serial line ppp poblem
On Sat, 13 Oct 2001, Eran Levy wrote: Aug 14 05:13:22 pppd[444]: pppd 2.3.11 started by root, uid 0 Aug 14 05:13:22 pppd[444]: Using interface ppp8 Aug 14 05:13:22 pppd[444]: Connect: ppp8 -- /dev/ttyS0 Aug 14 05:13:22 pppd[444]: ioctl(SIOCSIFMTU): No such device(19) Aug 14 05:13:22 pppd[444]: tcflush failed: Input/output error which pppd and which kernel? -- mulix http://www.advogato.com/person/mulix http://syscalltrack.sf.net/ = To unsubscribe, send mail to [EMAIL PROTECTED] with the word unsubscribe in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
