ipcop-devel Mailing List for IPCop Firewall

----- Original Message -----
From: "D.S.Hodgson (Administrator)" <AD...@ce...>
To: <ipc...@li...>
Sent: Sunday, November 30, 2003 1:58 PM
Subject: [IPCop-devel] build fail of adiusbadsl in 1.4 on RH 7.3


> Hi,
>
> I've been trying to build 1.4 on RH 7.3 and have hit a problem which
others seem to
> have had:
>
> in ipcop.build,log:
> ..
> Nov 29 18:55:13: Building Eagle driver USB ADSL
> <snip>
> make[1]: Entering directory
`/var/tmp/ipcop.build.6530/eagle-1.0.4/driver/user'
> gcc -O2 -pipe -Wall -pedantic -DLINUX -g adictrl.c -o adictrl
> In file included from adictrl.c:36:
> /usr/include/linux/usbdevice_fs.h:109: warning: ISO C forbids zero-size
array
> `iso_frame_desc'
> make[1]: Leaving directory
`/var/tmp/ipcop.build.6530/eagle-1.0.4/driver/user'

As I say before, I have exactly the same in my log and the warning is not
the problem.

But after is the real problem
> install: cannot create regular file
>
`/var/tmp/ipcop.build.6530/build/lib/modules/2.4.22/kernel/drivers/usb/adius
badsl.o':
> No such file or directory

You could change in make.sh like that

  make >> $LOGFILE 2>&1
  mkdir $TMPDIR/build/lib/modules/$KERNEL_VERSION/kernel/drivers/usb
  install -m 0664 -o root -g root adiusbadsl.o
$TMPDIR/build/lib/modules/$KERNEL_VERSION/kernel/drivers/usb/adiusbadsl.o >>
$LOGFILE 2>&1

this will create the usb directory
but normaly, this directory should be created before because there is other
drivers inside form the kernel (acm, catc,CDCether...).
So I suspect there is other problems during compilation of the other usb
drivers.
>
> and it's quite right, the directory mentioned doesn't exist, but the file
is to be found
> in /var/tmp/ipcop.build.6530/eagle-1.0.4/driver/adiusbadsl.o
>
> The link Giles gives doesn't seem to be available:
> http://fast800.tuxfamily.org/pub/article.php3?id_article=2
> can someone e-mail me the text?

It's specific for debian, so it's should not interest you on RH7.3

If this can help you

Gilles

Hi,

I've been trying to build 1.4 on RH 7.3 and have hit a problem which others seem to 
have had:

in ipcop.build,log:
..
Nov 29 18:55:13: Building Eagle driver USB ADSL
<snip>
make[1]: Entering directory `/var/tmp/ipcop.build.6530/eagle-1.0.4/driver/user'
gcc -O2 -pipe -Wall -pedantic -DLINUX -g adictrl.c -o adictrl
In file included from adictrl.c:36:
/usr/include/linux/usbdevice_fs.h:109: warning: ISO C forbids zero-size array 
`iso_frame_desc'
make[1]: Leaving directory `/var/tmp/ipcop.build.6530/eagle-1.0.4/driver/user'
install: cannot create regular file 
`/var/tmp/ipcop.build.6530/build/lib/modules/2.4.22/kernel/drivers/usb/adiusbadsl.o': 
No such file or directory

and it's quite right, the directory mentioned doesn't exist, but the file is to be found 
in /var/tmp/ipcop.build.6530/eagle-1.0.4/driver/adiusbadsl.o

Can someone forward me a fix as it seems to have happened before ...

It's mentioned here:
https://sourceforge.net/mailarchive/message.php?msg_id=6102940
and
https://sourceforge.net/mailarchive/message.php?msg_id=6043368

The link Giles gives doesn't seem to be available:
http://fast800.tuxfamily.org/pub/article.php3?id_article=2
can someone e-mail me the text?

Thanks
David

p.s. I have a fully patched and upgraded RH 7.3, with the packages mentioned on 
the build page checked and installed.  This instance is "clean" in that it has never 
been used for anything apart from building ipcop iso's.  As a result it has a fairly 
limited set of rpms present.  Just the core, gnome, some internet stuff, gcc 3.1 and 
stuff ipcop demands.  Again, if someone who has this working can send me their 
rpm list, I'll try augmenting my build to find the missing bit ...
====
D.S. Hodgson
Cerdo Ltd    EMail ad...@ce...

  PGP key  : CE 8B 73 4E BA 22 10 76
fingerprint: 32 73 C9 14 1A 43 37 62
=====
If it's PgP signed, it's official. If not ...

Rob Mattix wrote:
>> Ok, got it figured out.
>> 
>> For some reason, there was an extra comma in the
>> /var/ipcop/portfw/config file in the 12th line, just before the
>> comment. 
>> Don't know why it stopped loading at the 8th rule, but removed it
>> with vi and everything works like it's supposed to now.
>> 
>> Thanks for the help and sanity checks!!!
>> 
Great!

Another satisfied Ipcop customer - sort of close to 1 billion served :)

Darren

Ok, got it figured out.

For some reason, there was an extra comma in the /var/ipcop/portfw/config 
file in the 12th line, just before the comment.
Don't know why it stopped loading at the 8th rule, but removed it with vi 
and everything works like it's supposed to now.

Thanks for the help and sanity checks!!!

Rob

At 09:31 AM 11/28/2003, you wrote:

>Rob Mattix wrote:
> >> At 10:08 PM 11/27/2003, you wrote:
><SNIP>
> >>>
> >>> What exactly happens that you cannot enter more than 8, is there an
> >>> error code?
> >>> By chance are there any comma's in your comments that you are
> >>> adding to your port forwards - there is code in place to remove
> >>> comma's from the comment field, but it has been modified several
> >>> times in the last while and maybe it has introduced a bug.
> >>>
> >>> Darren
> >>> --__--__--
> >>
> >> It's not a problem entering them in the web gui, they don't show up
> >> when I look at the tables directly using the commands you list.
> >> There are not any comma's, however there are dashes but even the
> >> ones being inserted have them.
> >>
> >> It's just the first 8 rules listed in the gui that are being loaded.
> >>
> >> It's strange.  Hopefully, I'll get some time later to look deeper.
> >>
>Yes, it is strange. There people on the list with 20 or more rules,
>particularly people who are running voip, varous servers, etc.
>
>One of the rules must have something in it that breaks the setportfw.
>
>When you find time, go to the command line and do a /usr/local/bin/setportfw
>if there is an error it will usually complain about it and dump out of the
>program. This is all I can suggest on debugging at the moment.
>
>Darren
>
>
>---
>Incoming mail is certified Virus Free.
>Checked by AVG anti-virus system (http://www.grisoft.com).
>Version: 6.0.542 / Virus Database: 336 - Release Date: 11/18/2003

Rob Mattix wrote:
>> At 10:08 PM 11/27/2003, you wrote:
<SNIP>
>>>
>>> What exactly happens that you cannot enter more than 8, is there an
>>> error code?
>>> By chance are there any comma's in your comments that you are
>>> adding to your port forwards - there is code in place to remove
>>> comma's from the comment field, but it has been modified several
>>> times in the last while and maybe it has introduced a bug.
>>>
>>> Darren
>>> --__--__--
>>
>> It's not a problem entering them in the web gui, they don't show up
>> when I look at the tables directly using the commands you list.
>> There are not any comma's, however there are dashes but even the
>> ones being inserted have them.
>>
>> It's just the first 8 rules listed in the gui that are being loaded.
>>
>> It's strange.  Hopefully, I'll get some time later to look deeper.
>>
Yes, it is strange. There people on the list with 20 or more rules,
particularly people who are running voip, varous servers, etc.

One of the rules must have something in it that breaks the setportfw.

When you find time, go to the command line and do a /usr/local/bin/setportfw
if there is an error it will usually complain about it and dump out of the
program. This is all I can suggest on debugging at the moment.

Darren

At 10:08 PM 11/27/2003, you wrote:


>Rob Mattix wrote:
> >> I just recently noticed a problem with my IPCop box.
> >>
> >> It will only insert 8 rules in the PORTFWACCESS table.  This is on
> >> both
> >> IPCop 1.4 and 1.3 (with fixes 1-5).  Haven't had time to check when
> >> it stopped working or even if it is fix related.
> >>
> >> Just wanted to get this out there and see if anyone else has come
> >> across this.
> >>
>
>Must be a local problem as I have 12 port forward rules entered in my table,
>and all of them show up in the iptables chains:
>iptables -t nat -L PORTFW
>iptables -L PORTFWACCESS
>
>What exactly happens that you cannot enter more than 8, is there an error
>code?
>By chance are there any comma's in your comments that you are adding to your
>port forwards - there is code in place to remove comma's from the comment
>field, but it has been modified several times in the last while and maybe it
>has introduced a bug.
>
>Darren
>
>
>
>
>--__--__--

It's not a problem entering them in the web gui, they don't show up when I 
look at the tables directly using the commands you list.
There are not any comma's, however there are dashes but even the ones being 
inserted have them.

It's just the first 8 rules listed in the gui that are being loaded.

It's strange.  Hopefully, I'll get some time later to look deeper.

Thanks,
Rob

Rob Mattix wrote:
>> I just recently noticed a problem with my IPCop box.
>>
>> It will only insert 8 rules in the PORTFWACCESS table.  This is on
>> both
>> IPCop 1.4 and 1.3 (with fixes 1-5).  Haven't had time to check when
>> it stopped working or even if it is fix related.
>>
>> Just wanted to get this out there and see if anyone else has come
>> across this.
>>

Must be a local problem as I have 12 port forward rules entered in my table,
and all of them show up in the iptables chains:
iptables -t nat -L PORTFW
iptables -L PORTFWACCESS

What exactly happens that you cannot enter more than 8, is there an error
code?
By chance are there any comma's in your comments that you are adding to your
port forwards - there is code in place to remove comma's from the comment
field, but it has been modified several times in the last while and maybe it
has introduced a bug.

Darren

I just recently noticed a problem with my IPCop box.

It will only insert 8 rules in the PORTFWACCESS table.  This is on both 
IPCop 1.4 and 1.3 (with fixes 1-5).  Haven't had time to check when it 
stopped working or even if it is fix related.

Just wanted to get this out there and see if anyone else has come across this.

Oh and Happy Thanksgiving for those who celebrate today!

Robert

On Sat, 22 Nov 2003 10:45:40 +0100, 
"Gilles Espinasse" <gil...@wa...> wrote in message 
<01db01c3b0dd$6420ad70$01b5a8c0@PII350>:

> 
> ----- Original Message -----
> From: "Blademan" <bla...@ge...>
> To: "IPcop-devel" <ipc...@li...>
> Sent: Saturday, November 22, 2003 1:53 AM
> Subject: [IPCop-devel] Confirm location of makefile for IPCOP
> 
> 
> > I wanted to confirm that the location for the makefile for IPCOP is
> > located in cvs at:
> > /ipcop/ipcop/make.sh
> >
> > I'm gonna take a crack at making a Gentoo (http://www.gentoo.org)
> > ebuild. I'd like to see if we could build IPCOP for other arch such
> > as ppc and sparc. That way I can put my x86 P3 to better use. So I
> > wanted to confirm the loacation of the correct makefile before I get
> > knee deep into the process.
> >
> > TIA and Cheers for a great product,
> > Blademan
> > Gentoo Dev
> It's may be not an easy task. All .rpm are 386.rpm.

..well, all  these can be built from sources and the .spec files,
with 'rpmbuild -ta $sourceball ',  another way is go the .deb 
route, this would likely enable ipcop to support the same 
hardware as all Debian ports.

> Some usb adsl driver should compile for ppc, other not. I am not
> optimistic for sparc.
> Even if I know some of us try to use gentoo to compile IPCop for 386,
> I don't think someone succeed to run the result due to glibc errors.
> You may be interested by this post even if it's in french
> 
> http://www.ixus.net/modules.php?op=modload&name=Forum&file=viewtopic&topic=8
> 721&forum=10
> 
> Good luck!

..http://google.com/language_tools?hl=en ate above and produced:
http://translate.google.com/translate?u=http%3A%2F%2Fwww.ixus.net%2Fmodules.php%3Fop%3Dmodload%26name%3DForum%26file%3Dviewtopic%26topic%3D8721%26forum%3D10&langpair=fr%7Cen&hl=en&ie=ISO-8859-1&prev=%2Flanguage_tools

-- 
..med vennlig hilsen = with Kind Regards from Arnt... ;-)
...with a number of polar bear hunters in his ancestry...
  Scenarios always come in sets of three: 
  best case, worst case, and just in case.

----- Original Message -----
From: "Blademan" <bla...@ge...>
To: "IPcop-devel" <ipc...@li...>
Sent: Saturday, November 22, 2003 1:53 AM
Subject: [IPCop-devel] Confirm location of makefile for IPCOP


> I wanted to confirm that the location for the makefile for IPCOP is
> located in cvs at:
> /ipcop/ipcop/make.sh
>
> I'm gonna take a crack at making a Gentoo (http://www.gentoo.org)
> ebuild. I'd like to see if we could build IPCOP for other arch such as
> ppc and sparc. That way I can put my x86 P3 to better use. So I wanted
> to confirm the loacation of the correct makefile before I get knee deep
> into the process.
>
> TIA and Cheers for a great product,
> Blademan
> Gentoo Dev
It's may be not an easy task. All .rpm are 386.rpm.
Some usb adsl driver should compile for ppc, other not. I am not optimistic
for sparc.
Even if I know some of us try to use gentoo to compile IPCop for 386, I
don't think someone succeed to run the result due to glibc errors.
You may be interested by this post even if it's in french

http://www.ixus.net/modules.php?op=modload&name=Forum&file=viewtopic&topic=8
721&forum=10

Good luck!

I wanted to confirm that the location for the makefile for IPCOP is
located in cvs at:
/ipcop/ipcop/make.sh

I'm gonna take a crack at making a Gentoo (http://www.gentoo.org)
ebuild. I'd like to see if we could build IPCOP for other arch such as
ppc and sparc. That way I can put my x86 P3 to better use. So I wanted
to confirm the loacation of the correct makefile before I get knee deep
into the process.

TIA and Cheers for a great product,
Blademan
Gentoo Dev

----- Original Message -----
From: "Robert Kerr" <Lit...@xs...>
To: <ipc...@li...>
Sent: Thursday, November 20, 2003 4:09 PM
Subject: Re: [IPCop-devel] Should we not include netcard setup programs in
IPcop?


> In message <024101c3ad5f$094a79c0$01b5a8c0@PII350>
>           "Gilles Espinasse" <gil...@wa...> wrote:
>
> > I think to the programs on this page
http://www.scyld.com/diag/index.html
>
> > this package for pci cards
ftp://ftp.scyld.com/pub/diag/netdiag-2.4.tar.gz
> > and some others for ISA cards 3c5x9setup...
>
> > Because we have one more interface on V1.4, we will have more trouble to
> > let all cards work together.
>
> > It weight 840 kB for the pci-package. There is too a rpm available.
>
> The problem I see is where would we put them? You really need to use these
> programs before the install even starts or at least before the install
gets
> to the network card selection. For people doing an http/ftp install this
> means you'd have to somehow fit the programs on the boot floppy or driver
> floppy. I'm pretty sure neither of these have 840k to spare.
>
> Maybe I misunderstood?
>
 I know boot or driver floppy are full or almost full.
Even if it should be available for the green card, it's may only be needed
for the other cards.
So you may not need those program during install. You could always change
some settings you need after install.

If your green card work, you can do it remotly, but if it's work not, it is
easier to do that directly on IPcop if the setup program is include.

In my case, I had to change the IRQ on the second card after a nework
install. After the right program was found, compilated and copied on the
IPCop disk through SSH, it took only a few second. But find and compile the
setup program took more than a few second.
Because we accept more interfaces, we will see more conflict. So it's may be
fair to integrate the tools to solve the problems directly.
I know we can say : format a dos disk, take a setup for dos/win95 on the net
and reboot from the floppy

One of the common problems with ethernet modem is incorrect autonegociation
speed with RED card wich can be solved after installation.

----- Original Message -----
From: "David Jones" <dav...@po...>
To: <ipc...@li...>
Sent: Wednesday, November 19, 2003 11:40 AM
Subject: [IPCop-devel] IPCop problem v1.3.0


> Hello,
>
>
>
> I think i may have discovered a bug in this software, either that or
there's a problem with the PC it's running on.
>
>
>
> The Problem: After a few hours (2 - 3) whenever I try to access the
administration pages on the IPCop box, i get CGI errors, not sure of the
exact term's it stated, but it was along the lines of a CGI Script error.
>
>
>
> The PC:
>
>
>
> Pentium 133MHz
>
> 72MB RAM
>
> 1.0Gb Hard Disk

the most common error is due to a too big proxy cache size eating all the
memory with the table referencing all objects on cache.

On Wed, 19 Nov 2003 10:40:16 +0000, 
David Jones <dav...@po...> wrote in message 
<PM....@pm...>:

> Hello,
> 
> I think i may have discovered a bug in this software, either that or
> there's a problem with the PC it's running on.
> 
> The Problem: After a few hours (2 - 3) whenever I try to access the
> administration pages on the IPCop box, i get CGI errors, not sure of
> the exact term's it stated, but it was along the lines of a CGI Script
> error.
> 
> The PC: 
> Pentium 133MHz
> 72MB RAM
> 1.0Gb Hard Disk
> LiteOn CDROM
> Floppy Disk
> 3Com Etherlink 10/100
> Realtek 8139 100MBps
>
> I installed the OS with the floppy and a CDROM.
> 
> Not sure what is to blame...

..cut and paste and post the output as it appears to you, 
and we'll find out.

-- 
..med vennlig hilsen = with Kind Regards from Arnt... ;-)
...with a number of polar bear hunters in his ancestry...
  Scenarios always come in sets of three: 
  best case, worst case, and just in case.

On Thu, 20 Nov 2003 15:09:47 GMT, 
Robert Kerr <Lit...@xs...> wrote in message 
<1f4...@bl...>:

> In message <024101c3ad5f$094a79c0$01b5a8c0@PII350>
>           "Gilles Espinasse" <gil...@wa...> wrote:
> 
> > I think to the programs on this page
> > http://www.scyld.com/diag/index.html
> 
> > this package for pci cards
> > ftp://ftp.scyld.com/pub/diag/netdiag-2.4.tar.gz
> > and some others for ISA cards 3c5x9setup...
> 
> > Because we have one more interface on V1.4, we will have more
> > trouble to let all cards work together.
> 
> > It weight 840 kB for the pci-package. There is too a rpm available.
> 
> The problem I see is where would we put them? You really need to use
> these programs before the install even starts or at least before the
> install gets to the network card selection. For people doing an
> http/ftp install this means you'd have to somehow fit the programs on
> the boot floppy or driver floppy. 

..I disagree: people smart enough to get paid do remote installs over
internet, will also be smart enough to figure out how to swap nics etc 
after the initial and temporary install.  This can be done in a chrooted
install, or with a faked restore, or a combination, etc.  ;-)

..doing it "on desk", means you may have to kill a ssh session, 
move a net wire, and ssh in again.  No big deal.

-- 
..med vennlig hilsen = with Kind Regards from Arnt... ;-)
...with a number of polar bear hunters in his ancestry...
  Scenarios always come in sets of three: 
  best case, worst case, and just in case.

In message <024101c3ad5f$094a79c0$01b5a8c0@PII350>
          "Gilles Espinasse" <gil...@wa...> wrote:

> I think to the programs on this page http://www.scyld.com/diag/index.html

> this package for pci cards ftp://ftp.scyld.com/pub/diag/netdiag-2.4.tar.gz
> and some others for ISA cards 3c5x9setup...

> Because we have one more interface on V1.4, we will have more trouble to
> let all cards work together.

> It weight 840 kB for the pci-package. There is too a rpm available.

The problem I see is where would we put them? You really need to use these
programs before the install even starts or at least before the install gets
to the network card selection. For people doing an http/ftp install this
means you'd have to somehow fit the programs on the boot floppy or driver
floppy. I'm pretty sure neither of these have 840k to spare.

Maybe I misunderstood?

-- 
 Robert Kerr

Hello,



I think i may have discovered a bug in this software, either that or there's a problem with the PC it's running on.



The Problem: After a few hours (2 - 3) whenever I try to access the administration pages on the IPCop box, i get CGI errors, not sure of the exact term's it stated, but it was along the lines of a CGI Script error.



The PC: 



Pentium 133MHz

72MB RAM

1.0Gb Hard Disk

LiteOn CDROM

Floppy Disk

3Com Etherlink 10/100

Realtek 8139 100MBps



I installed the OS with the floppy and a CDROM.



Not sure what is to blame...



Thanks

David Jones


-- 
===========================

- - - - David Jones - - - -

===========================




___________________________________________________ 
Reduce your company's IT costs today with Officemaster. Sign up for a free trial!

http://www.officemaster.net

Melvin Backus escreveu:
>> Besides, those of us who use it at home and know what's going on on our
>> networks (either with or without the other user's knowledge) would 
>> probably
>> find it incredibly annoying if this were the default. I use non-standard
>> ports all the time for various things (including my own servers), and I
>> don't want to have to allow each one, or hack my 'tables rules to be 
>> able to
>> do so.
> 
> 
> How about something during the install which asks you for the desired 
> state.  Given that you've obviously seen the need for a firewall if 
> you're installing it, I'd say 2 or maybe 3 levels is more than adequate.

Hi folks.

Coincidentaly, I just have a talk over a Brazilian Linux site wich made 
a small evaluation of ipcop and this and the fact that ssh alows 
(actualy it's the only way it works now) was the week points detected.
I believe that an install option as stated above can be a very 
interesting compromise solution.

[]s

Adilson.

I think to the programs on this page http://www.scyld.com/diag/index.html

this package for pci cards ftp://ftp.scyld.com/pub/diag/netdiag-2.4.tar.gz
and some others for ISA cards 3c5x9setup...

Because we have one more interface on V1.4, we will have more trouble to let
all cards work together.

It weight 840 kB for the pci-package. There is too a rpm available.

Hello
Sorry to bother again, but I found another problem.
config.dat does not save the entered values anymore.
I found the following cause:
config.dat Line 128 reads:
<td align='center'><input type='submit' name='action' 
value='$tr{'save'}' /></td>

solution:
name='action' should be changed back to name='ACTION' so that it 
corresponds with line 23 $logsettings{'ACTION'} = '';.

Kind regards,

Roger

-- 
Roger Murer
E-Mail: r....@ti...

Hello,

Here I go again ;-).
I did a new build today and installed it using a backup from V1.3. I 
went trough the pages and changed some settings on the time.cgi page. By 
pressing the save button I got a server error page. I poked around a bit 
and saw that time.cgi was not able to write to settime.conf. This is due 
to that all the files in /var/ipcop/time/ have the ownership root:root, 
except for the file settings which has nobody:nobody.
This seems to only be a problem when using a backup to restore. A manual 
install (without backup) worked out of the box.

Kind regards,

Roger

-- 
Roger Murer
E-Mail: r....@ti...

Robert Kerr wrote:
> Oops, I know why that is, reminds me why people hate pointers in c. I'll
> commit a fix later today.

Hi

I did a rebuild and installed again, looks fine now.

Thank you

Roger

-- 
Roger Murer
E-Mail: r....@ti...

In message <3FB...@ti...>
          Roger Murer <r....@ti...> wrote:

> Hello,

> Against all advice, I am still playing around with 1.4a2 :-). With
> todays build (20031115) I noticed a strange problem with the installer.
> I am doing an install from scratch on a spare box, so I go through the
> whole setup.

Oops, I know why that is, reminds me why people hate pointers in c. I'll
commit a fix later today.

-- 
 Robert Kerr

Hello,

Against all advice, I am still playing around with 1.4a2 :-). With 
todays build (20031115) I noticed a strange problem with the installer. 
I am doing an install from scratch on a spare box, so I go through the 
whole setup.

I set green to 192.168.1.2/255.255.255.0 and do a install over http.
When I get to the screen "Enter the machine's host name" I get presented 
with ipc (funy normaly it would give me ipcop as default). I retype 
ipcop into the field and go on. I set the domain name to rmurer.ch and 
finish the setup with a network type of green + red (modem/isdn).
After the reboot, the hosts file looks like this:
127.0.0.1	localhost
192	ipc.rmu	ipc

After activating snort, a restartsnort gives me:
Bad GREEN_ADDRESS: 192

I did one more install (on an other box) using a V1.3 backup to restore 
the settings, but the problem is the same as when I do a manual install.

Kind regards,

Roger

-- 
Roger Murer
E-Mail: r....@ti...

----- Original Message -----
From: "Mark George Rainer" <mge...@ci...>
To: <ipc...@li...>
Cc: <mge...@ci...>
Sent: Thursday, November 13, 2003 2:21 PM
Subject: [IPCop-devel] Suspected bug/feature


> I don't know whether this is considered important but it concerned me for
> some time...
>
> AFAICT IPCop boots fine if there is an ADSL signal present but hangs
> otherwise, unless the USB modem (Speedtouch 330) is unplugged.
>
> I was trying to do as much testing as possible before my ADSL service was
> available.  Eventually I discovered that plugging in the modem after
> booting and then issuing the connect command would allow the firmware to
> load without hanging the OS.
>
> My concern relates to unattended recovery from power interruption.
>
>
I have reported that to the speedtouch mailing list in April before but I
was not able to change this behavior by myself.
Nobody care about it since yesterday when someone do a patch to add a
timeout on synchronisation so it's may be in the next version of modem_run.