[Bluez-users] Only one pand-daemon?

Mats Erik Andersson

author

2007-07-02 21:49 UTC #0 lore raw

Hello all,

I am facing the problem that I cannot start more than one listening pand-daemon, in spite of using two usb-dongles. The second call with 'pand --listen -e bdaddr ...' immediately ends with return value zero and no traces in any log file. I am running Bluez-utils 3.7 on Debian Etch. Slowly I am beginning to think that it could be DBUS that haunts me, but I have just begun studying the source code of pand to make sure. Any pointers are welcome.!

Best regards, Mats Erik Andersson

-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
Bluez-users mailing list [email protected]
https://lists.sourceforge.net/lists/listinfo/bluez-users

Marcel Holtmann

maintainer

2007-07-03 03:33 UTC #1 lore raw

Hi Mats,

Quoted 7 lines

I am facing the problem that I cannot start more than one listening pand-daemon, in spite of using two usb-dongles. The second call with 'pand --listen -e bdaddr ...' immediately ends with return value zero and no traces in any log file. I am running Bluez-utils 3.7 on Debian Etch. Slowly I am beginning to think that it could be DBUS that haunts me, but I have just begun studying the source code of pand to make sure.

I have no idea why you actually need two pand, but you can call pand with -i hciX to bind it to a specific adapter.

Regards

Marcel

-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
Bluez-users mailing list [email protected]
https://lists.sourceforge.net/lists/listinfo/bluez-users

Mats Erik Andersson

author

2007-07-03 11:39 UTC #2 lore raw

Dear Marcel and all other fellows,

tis 2007-07-03 klockan 05:33 +0200 skrev Marcel Holtmann:

Quoted 8 lines

Hi Mats,
Quoted 5 lines
I am facing the problem that I cannot start more than
one listening pand-daemon, in spite of using two usb-dongles. The second call with 'pand --listen -e bdaddr ...' immediately ends with return value zero and no traces in any log file.
I am running Bluez-utils 3.7 on Debiif test "$PAND_ENABLED" != "0"; then

Quoted 6 lines

an Etch. Slowly I am
Quoted 2 lines
beginning to think that it could be DBUS that haunts me, but I have just begun studying the source code of pand to make sure.
I have no idea why you actually need two pand, but you can call pand with -i hciX to bind it to a specific adapter.

I need more than one Bluetooth radio, i.e. usb-dongle, each providing a NAP service, since I am not content with only seven clients in one piconet. Hence my belief was that I could simply start one pand-daemon for each dongle and have my system up and running.

To that end I made an alteration to /etc/init.d/bluetooth to contain (within start_pand()) essentially

HCIDEVICES="0 1"
HCIBDADDR=(bdaddr0 bdaddr1) # in numerical values!

if test "$PAND_ENABLED" != "0"; then
for nr in $HCIDEVICES; do

    start-stop-daemon --start --quiet --exec $PAND_DAEMON -- \
         $PAND_OPTIONS --devup /etc/bluetooth/pan/dev-up \
         --ethernet "bnep$nr%d" --device ${HCIBDADDR[$nr]} \
         --pidfile /var/run/pand.${nr}.pid
    /usr/local/sbin/passkey-agent Buhh ${HCIBDADDR[$nr]} &

done
log_progress_msg "pand"
fi

The outcome is that only hci0 is serviced by pand and that it configures bnep00 at first call. There is no trace of a service for hci1, neither in the process list, in syslog, nor does it respond to "pand --connect bdadd1" from a remote machine.

Do you disagree to a need of servicing more than one piconet from a single machine, or is this not even implemented in the present day BlueZ-stack? Again, my server runs kernel 2.6.18 with Debian Etch and hence Bluez-utils 3.7. Have the later issues of Bluez-utils diversified in the above respect?

Best regards, Mats E A

-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
Bluez-users mailing list [email protected]
https://lists.sourceforge.net/lists/listinfo/bluez-users

Marcel Holtmann

maintainer

2007-07-04 01:48 UTC #3 lore raw

Hi Mats,

Quoted 10 lines

Quoted 10 lines
Quoted 5 lines
I am facing the problem that I cannot start more than
one listening pand-daemon, in spite of using two usb-dongles. The second call with 'pand --listen -e bdaddr ...' immediately ends with return value zero and no traces in any log file.
I am running Bluez-utils 3.7 on Debiif test "$PAND_ENABLED" != "0"; then
an Etch. Slowly I am
Quoted 2 lines
beginning to think that it could be DBUS that haunts me, but I have just begun studying the source code of pand to make sure.

Quoted 8 lines

Quoted 2 lines
I have no idea why you actually need two pand, but you can call pand with -i hciX to bind it to a specific adapter.
I need more than one Bluetooth radio, i.e. usb-dongle, each providing a NAP service, since I am not content with only seven clients in one piconet. Hence my belief was that I could simply start one pand-daemon for each dongle and have my system up and running.

by default pand listens on all attached dongles. So that should work out of the box.

Regards

Marcel

-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
Bluez-users mailing list [email protected]
https://lists.sourceforge.net/lists/listinfo/bluez-users

Mats Erik Andersson

author

2007-07-04 13:43 UTC #4 lore raw

[Bluez-users] dbus problem, Was: Only one pand-daemon?

Dear Marcel,

I did find one solution to this:

ons 2007-07-04 klockan 03:48 +0200 skrev Marcel Holtmann:

Quoted 9 lines

Hi Mats,
Quoted 7 lines
Quoted 6 lines
Quoted 5 lines
I am facing the problem that I cannot start more than one listening pand-daemon, in spite of using two usb-dongles.
beginning to think that it could be DBUS that haunts me, but I have just begun studying the source code of pand to make sure.

Quoted 8 lines

Quoted 4 lines
I need more than one Bluetooth radio, i.e. usb-dongle, each providing a NAP service, since I am not content with only seven clients in one piconet. Hence my belief was that I could simply start one pand-daemon for each dongle and have my system up and running.
by default pand listens on all attached dongles. So that should work out of the box.

I made an addition to the dbus-configuration in bluez-hcid.conf by inserting paths for hci0 and hci1 for root and console like this:

After that a single listening PAN daemon is receptive on both usb-dongles. My full configuration is attached here. It was already previously expanded, but it was the four lines with 'hci#' that brought the magic. I myself and probably other bluez-users, would improve our knowledge, could you be so kind as to commenting on what additions that could/should be eliminated from my configuration file. Please!

Best regards, Mats E A

</busconfig>

-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/

_______________________________________________
Bluez-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/bluez-users

Marcel Holtmann

maintainer

2007-07-05 08:10 UTC #5 lore raw

Re: [Bluez-users] dbus problem, Was: Only one pand-daemon?

Hi Mats,

Quoted 13 lines

Quoted 3 lines
by default pand listens on all attached dongles. So that should work out of the box.
I made an addition to the dbus-configuration in bluez-hcid.conf by inserting paths for hci0 and hci1 for root and console like this:
<policy user="root">
<allow own="org.bluez"/>
<allow send_path="/org/bluez/hci0"/>
<allow send_path="/org/bluez/hci1"/>
</policy>

Quoted 6 lines

After that a single listening PAN daemon is receptive on both usb-dongles. My full configuration is attached here. It was already previously expanded, but it was the four lines with 'hci#' that brought the magic. I myself and probably other bluez-users, would improve our knowledge, could you be so kind as to commenting on what additions that could/should be eliminated from my configuration file. Please!

that is non-sense. The pand is not interacting with D-Bus at all. Some other changes must have made this working for you.

Regards

Marcel

-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
Bluez-users mailing list [email protected]
https://lists.sourceforge.net/lists/listinfo/bluez-users

Mats Erik Andersson

author

2007-07-06 09:53 UTC #6 lore raw

Re: [Bluez-users] dbus problem for pand service

Hi Marcel,

tor 2007-07-05 klockan 10:10 +0200 skrev Marcel Holtmann:

Quoted 4 lines

Hi Mats,
=
Quoted 1 line
Quoted 1 line
by default pand listens on all attached dongles. So that should work =

out

Quoted 9 lines

Quoted 8 lines
Quoted 2 lines
of the box.
=
=
I made an addition to the dbus-configuration in bluez-hcid.conf by inserting paths for hci0 and hci1 for root and console like this:
=

Quoted 5 lines

Quoted 2 lines
After that a single listening PAN daemon is receptive on both usb-dongles.
that is non-sense. The pand is not interacting with D-Bus at all. Some other changes must have made this working for you.

a harsh word to call it 'non-sense' (zum st=FCtzig machen!). True, pand does not by itself call upon dbus, but the servicing daemon hcid does depend on dbus and all the more does passkey-agent for the authentications of pin codes depend in dbus. The standard code passkey-agent.c even contains

if (use_default) {
       path =3D "/org/bluez";
       method =3D "RegisterDefaultPasskeyAgent";
} else {
       path =3D "/org/bluez/hci0";
       method =3D "RegisterPasskeyAgent";
}

and the corresponding part for unregistering. It seems to me that =

thus it is absolutely essential to use 'passkey-agent -d OpenUp' to activate more than one dongle, since individual agents always tie themselfs to hci0. I do not understand all details yet, but the above code snippet shows that I will develop my own tailored passkey-agent for my own application. That is the natural state of affairs in our trade. Right! One always learn most by the mistakes and by scrutinizing our source code. =

Regards Mats E A

-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
Bluez-users mailing list [email protected]
https://lists.sourceforge.net/lists/listinfo/bluez-users

Marcel Holtmann

maintainer

2007-07-11 09:54 UTC #7 lore raw

Re: [Bluez-users] dbus problem for pand service

SGkgTWF0cywKCj4gPiA+ID4gYnkgZGVmYXVsdCBwYW5kIGxpc3RlbnMgb24gYWxsIGF0dGFjaGVk IGRvbmdsZXMuIFNvIHRoYXQgc2hvdWxkIHdvcmsgb3V0Cj4gPiA+ID4gb2YgdGhlIGJveC4KPiA+ ID4gPiAKPiA+ID4gCj4gPiA+IEkgbWFkZSBhbiBhZGRpdGlvbiB0byB0aGUgZGJ1cy1jb25maWd1 cmF0aW9uIGluIGJsdWV6LWhjaWQuY29uZiBieQo+ID4gPiBpbnNlcnRpbmcgcGF0aHMgZm9yIGhj aTAgYW5kIGhjaTEgZm9yIHJvb3QgYW5kIGNvbnNvbGUgbGlrZSB0aGlzOgo+ID4gPiAKPiA+ID4g QWZ0ZXIgdGhhdCBhIHNpbmdsZSBsaXN0ZW5pbmcgUEFOIGRhZW1vbiBpcyByZWNlcHRpdmUgb24g Ym90aAo+ID4gPiB1c2ItZG9uZ2xlcy4KPiAKPiA+IHRoYXQgaXMgbm9uLXNlbnNlLiBUaGUgcGFu ZCBpcyBub3QgaW50ZXJhY3Rpbmcgd2l0aCBELUJ1cyBhdCBhbGwuIFNvbWUKPiA+IG90aGVyIGNo YW5nZXMgbXVzdCBoYXZlIG1hZGUgdGhpcyB3b3JraW5nIGZvciB5b3UuCj4gCj4gCj4gYSBoYXJz aCB3b3JkIHRvIGNhbGwgaXQgJ25vbi1zZW5zZScgKHp1bSBzdMO8dHppZyBtYWNoZW4hKS4gVHJ1 ZSwgcGFuZAo+IGRvZXMgbm90IGJ5IGl0c2VsZiBjYWxsIHVwb24gZGJ1cywgYnV0IHRoZSBzZXJ2 aWNpbmcgZGFlbW9uIGhjaWQgZG9lcwo+IGRlcGVuZCBvbiBkYnVzIGFuZCBhbGwgdGhlIG1vcmUg ZG9lcyBwYXNza2V5LWFnZW50IGZvciB0aGUKPiBhdXRoZW50aWNhdGlvbnMgb2YgcGluIGNvZGVz IGRlcGVuZCBpbiBkYnVzLiBUaGUgc3RhbmRhcmQgY29kZQo+IHBhc3NrZXktYWdlbnQuYyBldmVu IGNvbnRhaW5zCj4gCj4gaWYgKHVzZV9kZWZhdWx0KSB7Cj4gICAgICAgIHBhdGggPSAiL29yZy9i bHVleiI7Cj4gICAgICAgIG1ldGhvZCA9ICJSZWdpc3RlckRlZmF1bHRQYXNza2V5QWdlbnQiOwo+ IH0gZWxzZSB7Cj4gICAgICAgIHBhdGggPSAiL29yZy9ibHVlei9oY2kwIjsKPiAgICAgICAgbWV0 aG9kID0gIlJlZ2lzdGVyUGFzc2tleUFnZW50IjsKPiB9Cj4gCj4gYW5kIHRoZSBjb3JyZXNwb25k aW5nIHBhcnQgZm9yIHVucmVnaXN0ZXJpbmcuIEl0IHNlZW1zIHRvIG1lIHRoYXQgCj4gdGh1cyBp dCBpcyBhYnNvbHV0ZWx5IGVzc2VudGlhbCB0byB1c2UgJ3Bhc3NrZXktYWdlbnQgLWQgT3BlblVw Jwo+IHRvIGFjdGl2YXRlIG1vcmUgdGhhbiBvbmUgZG9uZ2xlLCBzaW5jZSBpbmRpdmlkdWFsIGFn ZW50cyBhbHdheXMKPiB0aWUgdGhlbXNlbGZzIHRvIGhjaTAuIEkgZG8gbm90IHVuZGVyc3RhbmQg YWxsIGRldGFpbHMgeWV0LCBidXQgdGhlCj4gYWJvdmUgY29kZSBzbmlwcGV0IHNob3dzIHRoYXQg SSB3aWxsIGRldmVsb3AgbXkgb3duIHRhaWxvcmVkCj4gcGFzc2tleS1hZ2VudCBmb3IgbXkgb3du IGFwcGxpY2F0aW9uLiBUaGF0IGlzIHRoZSBuYXR1cmFsIHN0YXRlIG9mCj4gYWZmYWlycyBpbiBv dXIgdHJhZGUuIFJpZ2h0ISBPbmUgYWx3YXlzIGxlYXJuIG1vc3QgYnkgdGhlIG1pc3Rha2VzCj4g YW5kIGJ5IHNjcnV0aW5pemluZyBvdXIgc291cmNlIGNvZGUuIAoKdGhpcyBpcyB3aHkgd2UgaGF2 ZSB0aGUgcGFzc2tleSBhZ2VudCBpbnRlcmZhY2UuIEl0IGlzIG1lYW50IHRvIGJlIHRoYXQKZXZl cnlib2R5IGNhbiBkZXZlbG9wIHRoZWlyIG93biBwYXNza2V5IGhhbmRsZXJzIGZvciB0aGVpciBu ZWVkcy4KClJlZ2FyZHMKCk1hcmNlbAoKCgotLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0t LS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tClRoaXMgU0YubmV0IGVt YWlsIGlzIHNwb25zb3JlZCBieSBEQjIgRXhwcmVzcwpEb3dubG9hZCBEQjIgRXhwcmVzcyBDIC0g dGhlIEZSRUUgdmVyc2lvbiBvZiBEQjIgZXhwcmVzcyBhbmQgdGFrZQpjb250cm9sIG9mIHlvdXIg WE1MLiBObyBsaW1pdHMuIEp1c3QgZGF0YS4gQ2xpY2sgdG8gZ2V0IGl0IG5vdy4KaHR0cDovL3Nv dXJjZWZvcmdlLm5ldC9wb3dlcmJhci9kYjIvCl9fX19fX19fX19fX19fX19fX19fX19fX19fX19f X19fX19fX19fX19fX19fX19fCkJsdWV6LXVzZXJzIG1haWxpbmcgbGlzdApCbHVlei11c2Vyc0Bs aXN0cy5zb3VyY2Vmb3JnZS5uZXQKaHR0cHM6Ly9saXN0cy5zb3VyY2Vmb3JnZS5uZXQvbGlzdHMv bGlzdGluZm8vYmx1ZXotdXNlcnMK