[Bluez-devel] piconet..

[Bluez-devel] piconet..

[Steffen Larsen]

[-- Attachment #1: Type: text/plain, Size: 996 bytes --]

Hi again!,

I am trying to form a piconet, consisting of one server and two clients. I have snaped the most of the code from l2test.c and also made an advertising service and all. My problem is just that one one client can connect and the other one fails (host is down, i belive) . I have implemented the server using fork(), so that the parent process is still listening for new connections and closing my sockets correct (as what I can see..).
I have browsed this newsgroup for answers, and can see something about setting the master/slave switch should help. I have also tried this, but it dosn't help. Anyone that have a clue about, what could be wrong?? 

PS. I belive that the PSM in the L2CAP layer can serve more than one client or am I totally wrong? otherwise I should update the PSM everytime a new client connect..

I anyone want the source code, I can send it. But I haven't attached it in this document, due to the size of it.. :-)


-Thanks in advance!

/Steffen Larsen

[-- Attachment #2: Type: text/html, Size: 1842 bytes --]

Re: [Bluez-devel] piconet..

[Marcel Holtmann]

Hi Steffen,

> I am trying to form a piconet, consisting of one server and two
> clients. I have snaped the most of the code from l2test.c and also
> made an advertising service and all. My problem is just that one one
> client can connect and the other one fails (host is down, i belive) .
> I have implemented the server using fork(), so that the parent process
> is still listening for new connections and closing my sockets correct
> (as what I can see..).
> I have browsed this newsgroup for answers, and can see something about
> setting the master/slave switch should help. I have also tried this,
> but it dosn't help. Anyone that have a clue about, what could be
> wrong?? 

show us the output of "hciconfig -a" for your devices.
 
> PS. I belive that the PSM in the L2CAP layer can serve more than one
> client or am I totally wrong? otherwise I should update the PSM
> everytime a new client connect..

It can also serve multiple PSM on the same ACL link.

Regards

Marcel




-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/
_______________________________________________
Bluez-devel mailing list
Bluez-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bluez-devel

Re: [Bluez-devel] piconet..

[Steffen Larsen]

Hi Marcel,

> Hi Steffen,
> 
> > I am trying to form a piconet, consisting of one server and two
> > clients. I have snaped the most of the code from l2test.c and also
> > made an advertising service and all. My problem is just that one one
> > client can connect and the other one fails (host is down, i belive) .
> > I have implemented the server using fork(), so that the parent process
> > is still listening for new connections and closing my sockets correct
> > (as what I can see..).
> > I have browsed this newsgroup for answers, and can see something about
> > setting the master/slave switch should help. I have also tried this,
> > but it dosn't help. Anyone that have a clue about, what could be
> > wrong?? 
> 
> show us the output of "hciconfig -a" for your devices.

ok.. I will do this later today. :-)
  
> > PS. I belive that the PSM in the L2CAP layer can serve more than one
> > client or am I totally wrong? otherwise I should update the PSM
> > everytime a new client connect..
> 
> It can also serve multiple PSM on the same ACL link.

Ok.. how do I do that?

Currently I am setting the PSM on my socket like this:

<code snip>
loc_addr.l2_family = AF_BLUETOOTH;

bacpy(&loc_addr.l2_bdaddr, local_bdaddr);

loc_addr.l2_psm = htobs(psm);


if (bind(sk, (struct sockaddr *) &loc_addr, sizeof(loc_addr)) < 0) {

printf("could not bind socket: %i\n", sk);

syslog(LOG_ERR, "Can't bind socket. %s(%d)", strerror(errno), errno);

exit(1);

}

</code snip>


/Steffen Larsen


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/
_______________________________________________
Bluez-devel mailing list
Bluez-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bluez-devel

Re: [Bluez-devel] piconet..

[Marcel Holtmann]

Hi Steffen,

> > > PS. I belive that the PSM in the L2CAP layer can serve more than one
> > > client or am I totally wrong? otherwise I should update the PSM
> > > everytime a new client connect..
> > 
> > It can also serve multiple PSM on the same ACL link.
> 
> Ok.. how do I do that?
> 
> Currently I am setting the PSM on my socket like this:
> 
> <code snip>
> loc_addr.l2_family = AF_BLUETOOTH;
> 
> bacpy(&loc_addr.l2_bdaddr, local_bdaddr);
> 
> loc_addr.l2_psm = htobs(psm);
> 
> 
> if (bind(sk, (struct sockaddr *) &loc_addr, sizeof(loc_addr)) < 0) {
> 
> printf("could not bind socket: %i\n", sk);
> 
> syslog(LOG_ERR, "Can't bind socket. %s(%d)", strerror(errno), errno);
> 
> exit(1);
> 
> }
> 
> </code snip>

you call listen() and then accept() for every incoming connection. It is
the same as for TCP/IP.

Regards

Marcel




-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/
_______________________________________________
Bluez-devel mailing list
Bluez-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bluez-devel

Re: [Bluez-devel] piconet..

[Steffen Larsen]

Hi Marcel,

> > > > PS. I belive that the PSM in the L2CAP layer can serve more than one
> > > > client or am I totally wrong? otherwise I should update the PSM
> > > > everytime a new client connect..
> > >
> > > It can also serve multiple PSM on the same ACL link.
> >
> > Ok.. how do I do that?
> >

> you call listen() and then accept() for every incoming connection. It is
> the same as for TCP/IP.

Ok.. I am already doing this in my parent thread. So the problem must be
somewhere else.
I'll try to do the hciconfig -a in a brief moment and sent the output to the
list... :-)

/Steffen Larsen



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/
_______________________________________________
Bluez-devel mailing list
Bluez-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bluez-devel

Re: [Bluez-devel] piconet..

[Steffen Larsen]

Hi Marcel and others,  ;-)

> > > > > PS. I belive that the PSM in the L2CAP layer can serve more than
one
> > > > > client or am I totally wrong? otherwise I should update the PSM
> > > > > everytime a new client connect..
> > > >
> > > > It can also serve multiple PSM on the same ACL link.
> > >
> > > Ok.. how do I do that?
> > >
>
> > you call listen() and then accept() for every incoming connection. It is
> > the same as for TCP/IP.
>
> Ok.. I am already doing this in my parent thread. So the problem must be
> somewhere else.
> I'll try to do the hciconfig -a in a brief moment and sent the output to
the
> list... :-)

My source code for most of my BTserver can be found at:
http://www.diku.dk/hjemmesider/studerende/zool/bt/prot.cc
As you can see in my souce (see the listenL2CAP method), I do almost as you
told me to do and what l2test.c is doing.
I use the standard bind, listen and accept methods and spawns of a new
child, when I get a connection. But I still can't make it work when I have
more than one client (a piconet). Why is that? can you see a minor flaw in
my code?

I have made a hciconfig -a with one of my USB dongles (it is a DBT 120
unit):

hci0: Type: USB
 BD Address: 00:40:05:60:41:0D ACL MTU: 192:8  SCO MTU: 64:8
 UP RUNNING PSCAN ISCAN
 RX bytes:653 acl:0 sco:0 events:19 errors:0
 TX bytes:314 acl:0 sco:0 commands:18 errors:0
 Features: 0xff 0xff 0x0f 0x00 0x00 0x00 0x00 0x00
 Packet type: DM1 DM3 DM5 DH1 DH3 DH5 HV1 HV2 HV3
 Link policy: RSWITCH HOLD SNIFF PARK
 Link mode: SLAVE ACCEPT
 Name: 'fsck-0'
 Class: 0x000100
 Service Classes: Unspecified
 Device Class: Computer, Uncategorized
 HCI Ver: 1.1 (0x1) HCI Rev: 0x1bb LMP Ver: 1.1 (0x1) LMP Subver: 0x1bb
 Manufacturer: Cambridge Silicon Radio (10)

I can only give you this one, because one of my friends have the other two
dongles. I will though sent the output later tonight, when I get them back.
:-)

PS: It also seems like that the l2test works with more clients.. I really
can't see the big difference in my code and l2test.. But mabye it is me..
;-)


/Steffen Larsen



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/
_______________________________________________
Bluez-devel mailing list
Bluez-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bluez-devel

Re: [Bluez-devel] piconet..

[Marcel Holtmann]

Hi Steffen,

> My source code for most of my BTserver can be found at:
> http://www.diku.dk/hjemmesider/studerende/zool/bt/prot.cc
> As you can see in my souce (see the listenL2CAP method), I do almost as you
> told me to do and what l2test.c is doing.
> I use the standard bind, listen and accept methods and spawns of a new
> child, when I get a connection. But I still can't make it work when I have
> more than one client (a piconet). Why is that? can you see a minor flaw in
> my code?
> 
> I have made a hciconfig -a with one of my USB dongles (it is a DBT 120
> unit):
> 
> hci0: Type: USB
>  BD Address: 00:40:05:60:41:0D ACL MTU: 192:8  SCO MTU: 64:8
>  UP RUNNING PSCAN ISCAN
>  RX bytes:653 acl:0 sco:0 events:19 errors:0
>  TX bytes:314 acl:0 sco:0 commands:18 errors:0
>  Features: 0xff 0xff 0x0f 0x00 0x00 0x00 0x00 0x00
>  Packet type: DM1 DM3 DM5 DH1 DH3 DH5 HV1 HV2 HV3
>  Link policy: RSWITCH HOLD SNIFF PARK
>  Link mode: SLAVE ACCEPT
>  Name: 'fsck-0'
>  Class: 0x000100
>  Service Classes: Unspecified
>  Device Class: Computer, Uncategorized
>  HCI Ver: 1.1 (0x1) HCI Rev: 0x1bb LMP Ver: 1.1 (0x1) LMP Subver: 0x1bb
>  Manufacturer: Cambridge Silicon Radio (10)
> 
> I can only give you this one, because one of my friends have the other two
> dongles. I will though sent the output later tonight, when I get them back.
> :-)

so far so good, but check with "hciconfig hci0 revision" for the
firmware version. This one is too old. It should be HCI 16.4 or newer,
but if this is a DBT-120 Rev. B3 you might have a chance to update it.

> PS: It also seems like that the l2test works with more clients.. I really
> can't see the big difference in my code and l2test.. But mabye it is me..
> ;-)

I think you should use fork(). What is wrong with using poll() or
select() and an event loop.

And btw since you use code from l2test.c etc. you should include the GPL
license statement at the top of your code.

Regards

Marcel




-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/
_______________________________________________
Bluez-devel mailing list
Bluez-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bluez-devel

Re: [Bluez-devel] piconet..

[Steffen Larsen]

Hi again Marcel,

> > My source code for most of my BTserver can be found at:
> > http://www.diku.dk/hjemmesider/studerende/zool/bt/prot.cc
> > As you can see in my souce (see the listenL2CAP method), I do almost as
you
> > told me to do and what l2test.c is doing.
> > I use the standard bind, listen and accept methods and spawns of a new
> > child, when I get a connection. But I still can't make it work when I
have
> > more than one client (a piconet). Why is that? can you see a minor flaw
in
> > my code?
> >
> > I have made a hciconfig -a with one of my USB dongles (it is a DBT 120
> > unit):
> >
> > hci0: Type: USB
> >  BD Address: 00:40:05:60:41:0D ACL MTU: 192:8  SCO MTU: 64:8
> >  UP RUNNING PSCAN ISCAN
> >  RX bytes:653 acl:0 sco:0 events:19 errors:0
> >  TX bytes:314 acl:0 sco:0 commands:18 errors:0
> >  Features: 0xff 0xff 0x0f 0x00 0x00 0x00 0x00 0x00
> >  Packet type: DM1 DM3 DM5 DH1 DH3 DH5 HV1 HV2 HV3
> >  Link policy: RSWITCH HOLD SNIFF PARK
> >  Link mode: SLAVE ACCEPT
> >  Name: 'fsck-0'
> >  Class: 0x000100
> >  Service Classes: Unspecified
> >  Device Class: Computer, Uncategorized
> >  HCI Ver: 1.1 (0x1) HCI Rev: 0x1bb LMP Ver: 1.1 (0x1) LMP Subver: 0x1bb
> >  Manufacturer: Cambridge Silicon Radio (10)
> >
> > I can only give you this one, because one of my friends have the other
two
> > dongles. I will though sent the output later tonight, when I get them
back.
> > :-)

I got the two others:

 hci0: Type: USB
BD Address: 00:01:53:00:14:EE ACL MTU: 192:8 SCO MTU: 64:8
UP RUNNING PSCAN ISCAN
RX bytes:81 acl:0 sco:0 events:10 errors:0
TX bytes:39 acl:0 sco:0 commands:10 errors:0
Features: 0xff 0xff 0x0f 0x00 0x00 0x00 0x00 0x00
Packet type: DM1 DM3 DM5 DH1 DH3 DH5 HV1 HV2 HV3
Link policy:
Link mode: SLAVE ACCEPT
Name: 'knoglen'
Class: 0x000000
Service Classes: Unspecified
Device Class: Miscellaneous,
HCI Ver: 1.1 (0x1) HCI Rev: 0x20d LMP Ver: 1.1 (0x1) LMP Subver: 0x20d
Manufacturer: Cambridge Silicon Radio (10)

hci0: Type: USB
BD Address: 00:01:53:00:0D:B5 ACL MTU: 192:8 SCO MTU: 64:8
UP RUNNING PSCAN ISCAN
RX bytes:87 acl:0 sco:0 events:11 errors:0
TX bytes:39 acl:0 sco:0 commands:10 errors:0
Features: 0xff 0xff 0x0f 0x00 0x00 0x00 0x00 0x00
Packet type: DM1 DM3 DM5 DH1 DH3 DH5 HV1 HV2 HV3
Link policy:
Link mode: SLAVE ACCEPT
Name: 'JESPER'
Class: 0x000000
Service Classes: Unspecified
Device Class: Miscellaneous,
HCI Ver: 1.1 (0x1) HCI Rev: 0x20d LMP Ver: 1.1 (0x1) LMP Subver: 0x20d
Manufacturer: Cambridge Silicon Radio (10)


> so far so good, but check with "hciconfig hci0 revision" for the
> firmware version. This one is too old. It should be HCI 16.4 or newer,
> but if this is a DBT-120 Rev. B3 you might have a chance to update it.

Well. It is the B3 revision, so I would be able to do that. :-)

> I think you should use fork(). What is wrong with using poll() or
> select() and an event loop.

Hmm.. This dosn't really answer my question. I am using fork()., and would
like to get my application to work. I tried to compare it towards the
l2test, but couldn't see that much difference. And the l2test works with the
same dongles; connecting two clients against one server. Can you point
somewhere in my code, where I should do some changes, so it could work??

I really starts to nag me, why my code dosn't work. All things works fine
(like connecting/disconnecting etc.) with one clients, but with two... :-(
And I have tried to make my server MASTER by setting the L2CAP_LM_MASTER in
the link manager.

> And btw since you use code from l2test.c etc. you should include the GPL
> license statement at the top of your code.

No worry.. I will.

-Thanks again for a quick answer! :-)

/Steffen Larsen



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/
_______________________________________________
Bluez-devel mailing list
Bluez-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bluez-devel

Re: [Bluez-devel] piconet..

[Marcel Holtmann]

Hi Steffen,

> I got the two others:
> 
>  hci0: Type: USB
> BD Address: 00:01:53:00:14:EE ACL MTU: 192:8 SCO MTU: 64:8
> UP RUNNING PSCAN ISCAN
> RX bytes:81 acl:0 sco:0 events:10 errors:0
> TX bytes:39 acl:0 sco:0 commands:10 errors:0
> Features: 0xff 0xff 0x0f 0x00 0x00 0x00 0x00 0x00
> Packet type: DM1 DM3 DM5 DH1 DH3 DH5 HV1 HV2 HV3
> Link policy:
> Link mode: SLAVE ACCEPT
> Name: 'knoglen'
> Class: 0x000000
> Service Classes: Unspecified
> Device Class: Miscellaneous,
> HCI Ver: 1.1 (0x1) HCI Rev: 0x20d LMP Ver: 1.1 (0x1) LMP Subver: 0x20d
> Manufacturer: Cambridge Silicon Radio (10)
> 
> hci0: Type: USB
> BD Address: 00:01:53:00:0D:B5 ACL MTU: 192:8 SCO MTU: 64:8
> UP RUNNING PSCAN ISCAN
> RX bytes:87 acl:0 sco:0 events:11 errors:0
> TX bytes:39 acl:0 sco:0 commands:10 errors:0
> Features: 0xff 0xff 0x0f 0x00 0x00 0x00 0x00 0x00
> Packet type: DM1 DM3 DM5 DH1 DH3 DH5 HV1 HV2 HV3
> Link policy:
> Link mode: SLAVE ACCEPT
> Name: 'JESPER'
> Class: 0x000000
> Service Classes: Unspecified
> Device Class: Miscellaneous,
> HCI Ver: 1.1 (0x1) HCI Rev: 0x20d LMP Ver: 1.1 (0x1) LMP Subver: 0x20d
> Manufacturer: Cambridge Silicon Radio (10)

these two are HCI 16.4 and so there should be no problem. See my CSR
page why it is good to have HCI 16.x at least.

> > so far so good, but check with "hciconfig hci0 revision" for the
> > firmware version. This one is too old. It should be HCI 16.4 or newer,
> > but if this is a DBT-120 Rev. B3 you might have a chance to update it.
> 
> Well. It is the B3 revision, so I would be able to do that. :-)

I suppose you do that (at your own risk of course).

> Hmm.. This dosn't really answer my question. I am using fork()., and would
> like to get my application to work. I tried to compare it towards the
> l2test, but couldn't see that much difference. And the l2test works with the
> same dongles; connecting two clients against one server. Can you point
> somewhere in my code, where I should do some changes, so it could work??

Getting the fork stuff right is not very easy. You must carefully think
about what file descriptors to close at what time and in what process. I
am not going to check your code for such mistakes, because it will take
me too much time and the other reason is that it is C++ and not C.

> I really starts to nag me, why my code dosn't work. All things works fine
> (like connecting/disconnecting etc.) with one clients, but with two... :-(
> And I have tried to make my server MASTER by setting the L2CAP_LM_MASTER in
> the link manager.

Update the dongle or try to swap the pre HCI 16.4 one. The master option
can help, but you must choose the master device carefully.

Regards

Marcel




-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/
_______________________________________________
Bluez-devel mailing list
Bluez-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bluez-devel

Re: [Bluez-devel] piconet..

[Steffen Larsen]

Hi Marcel,

> > I got the two others:
> >
> >  hci0: Type: USB
> > BD Address: 00:01:53:00:14:EE ACL MTU: 192:8 SCO MTU: 64:8
> > UP RUNNING PSCAN ISCAN
> > RX bytes:81 acl:0 sco:0 events:10 errors:0
> > TX bytes:39 acl:0 sco:0 commands:10 errors:0
> > Features: 0xff 0xff 0x0f 0x00 0x00 0x00 0x00 0x00
> > Packet type: DM1 DM3 DM5 DH1 DH3 DH5 HV1 HV2 HV3
> > Link policy:
> > Link mode: SLAVE ACCEPT
> > Name: 'knoglen'
> > Class: 0x000000
> > Service Classes: Unspecified
> > Device Class: Miscellaneous,
> > HCI Ver: 1.1 (0x1) HCI Rev: 0x20d LMP Ver: 1.1 (0x1) LMP Subver: 0x20d
> > Manufacturer: Cambridge Silicon Radio (10)
> >
> > hci0: Type: USB
> > BD Address: 00:01:53:00:0D:B5 ACL MTU: 192:8 SCO MTU: 64:8
> > UP RUNNING PSCAN ISCAN
> > RX bytes:87 acl:0 sco:0 events:11 errors:0
> > TX bytes:39 acl:0 sco:0 commands:10 errors:0
> > Features: 0xff 0xff 0x0f 0x00 0x00 0x00 0x00 0x00
> > Packet type: DM1 DM3 DM5 DH1 DH3 DH5 HV1 HV2 HV3
> > Link policy:
> > Link mode: SLAVE ACCEPT
> > Name: 'JESPER'
> > Class: 0x000000
> > Service Classes: Unspecified
> > Device Class: Miscellaneous,
> > HCI Ver: 1.1 (0x1) HCI Rev: 0x20d LMP Ver: 1.1 (0x1) LMP Subver: 0x20d
> > Manufacturer: Cambridge Silicon Radio (10)
>
> these two are HCI 16.4 and so there should be no problem. See my CSR
> page why it is good to have HCI 16.x at least.
>
> > > so far so good, but check with "hciconfig hci0 revision" for the
> > > firmware version. This one is too old. It should be HCI 16.4 or newer,
> > > but if this is a DBT-120 Rev. B3 you might have a chance to update it.
> >
> > Well. It is the B3 revision, so I would be able to do that. :-)
>
> I suppose you do that (at your own risk of course).
>
> > Hmm.. This dosn't really answer my question. I am using fork()., and
would
> > like to get my application to work. I tried to compare it towards the
> > l2test, but couldn't see that much difference. And the l2test works with
the
> > same dongles; connecting two clients against one server. Can you point
> > somewhere in my code, where I should do some changes, so it could work??
>
> Getting the fork stuff right is not very easy. You must carefully think
> about what file descriptors to close at what time and in what process. I
> am not going to check your code for such mistakes, because it will take
> me too much time and the other reason is that it is C++ and not C.

I understand. I have been spending alot of time in it my self. I will look
deeper into the socket, and which of them that have been closed and at which
time... But I think I have checked this a million times.. But I will try
again! :-)

> Update the dongle or try to swap the pre HCI 16.4 one. The master option
> can help, but you must choose the master device carefully.

Yes I can try to swap the devices. You mention that the master device should
be carefully choosen. I what way?
I have just assumed that my server should be the master, and thats that.
What should I taken in to consideration?

-Once again thanks for all the quick answers! :-)

/Steffen Larsen



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/
_______________________________________________
Bluez-devel mailing list
Bluez-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bluez-devel

Re: [Bluez-devel] piconet..

[Steffen Larsen]

HI Marcel and other!,

I found out what I did wrong. All my sockets and fork()'ing just worked
fine. My problem was that I tried to make:


----- Original Message -----
From: "Steffen Larsen" <zool@zool.dk>
To: <bluez-devel@lists.sourceforge.net>
Sent: Saturday, December 11, 2004 6:50 PM
Subject: Re: [Bluez-devel] piconet..


> Hi Marcel,
>
> > > I got the two others:
> > >
> > >  hci0: Type: USB
> > > BD Address: 00:01:53:00:14:EE ACL MTU: 192:8 SCO MTU: 64:8
> > > UP RUNNING PSCAN ISCAN
> > > RX bytes:81 acl:0 sco:0 events:10 errors:0
> > > TX bytes:39 acl:0 sco:0 commands:10 errors:0
> > > Features: 0xff 0xff 0x0f 0x00 0x00 0x00 0x00 0x00
> > > Packet type: DM1 DM3 DM5 DH1 DH3 DH5 HV1 HV2 HV3
> > > Link policy:
> > > Link mode: SLAVE ACCEPT
> > > Name: 'knoglen'
> > > Class: 0x000000
> > > Service Classes: Unspecified
> > > Device Class: Miscellaneous,
> > > HCI Ver: 1.1 (0x1) HCI Rev: 0x20d LMP Ver: 1.1 (0x1) LMP Subver: 0x20d
> > > Manufacturer: Cambridge Silicon Radio (10)
> > >
> > > hci0: Type: USB
> > > BD Address: 00:01:53:00:0D:B5 ACL MTU: 192:8 SCO MTU: 64:8
> > > UP RUNNING PSCAN ISCAN
> > > RX bytes:87 acl:0 sco:0 events:11 errors:0
> > > TX bytes:39 acl:0 sco:0 commands:10 errors:0
> > > Features: 0xff 0xff 0x0f 0x00 0x00 0x00 0x00 0x00
> > > Packet type: DM1 DM3 DM5 DH1 DH3 DH5 HV1 HV2 HV3
> > > Link policy:
> > > Link mode: SLAVE ACCEPT
> > > Name: 'JESPER'
> > > Class: 0x000000
> > > Service Classes: Unspecified
> > > Device Class: Miscellaneous,
> > > HCI Ver: 1.1 (0x1) HCI Rev: 0x20d LMP Ver: 1.1 (0x1) LMP Subver: 0x20d
> > > Manufacturer: Cambridge Silicon Radio (10)
> >
> > these two are HCI 16.4 and so there should be no problem. See my CSR
> > page why it is good to have HCI 16.x at least.
> >
> > > > so far so good, but check with "hciconfig hci0 revision" for the
> > > > firmware version. This one is too old. It should be HCI 16.4 or
newer,
> > > > but if this is a DBT-120 Rev. B3 you might have a chance to update
it.
> > >
> > > Well. It is the B3 revision, so I would be able to do that. :-)
> >
> > I suppose you do that (at your own risk of course).
> >
> > > Hmm.. This dosn't really answer my question. I am using fork()., and
> would
> > > like to get my application to work. I tried to compare it towards the
> > > l2test, but couldn't see that much difference. And the l2test works
with
> the
> > > same dongles; connecting two clients against one server. Can you point
> > > somewhere in my code, where I should do some changes, so it could
work??
> >
> > Getting the fork stuff right is not very easy. You must carefully think
> > about what file descriptors to close at what time and in what process. I
> > am not going to check your code for such mistakes, because it will take
> > me too much time and the other reason is that it is C++ and not C.
>
> I understand. I have been spending alot of time in it my self. I will look
> deeper into the socket, and which of them that have been closed and at
which
> time... But I think I have checked this a million times.. But I will try
> again! :-)
>
> > Update the dongle or try to swap the pre HCI 16.4 one. The master option
> > can help, but you must choose the master device carefully.
>
> Yes I can try to swap the devices. You mention that the master device
should
> be carefully choosen. I what way?
> I have just assumed that my server should be the master, and thats that.
> What should I taken in to consideration?
>
> -Once again thanks for all the quick answers! :-)
>
> /Steffen Larsen
>
>
>
> -------------------------------------------------------
> SF email is sponsored by - The IT Product Guide
> Read honest & candid reviews on hundreds of IT Products from real users.
> Discover which products truly live up to the hype. Start reading now.
> http://productguide.itmanagersjournal.com/
> _______________________________________________
> Bluez-devel mailing list
> Bluez-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/bluez-devel
>



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/
_______________________________________________
Bluez-devel mailing list
Bluez-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bluez-devel

Re: [Bluez-devel] piconet..

[Steffen Larsen]

Sorry!.. Was to quick on the 'send' button!

My problem was that I used the:

bacpy(&loc_addr.l2_bdaddr, local_bdaddr);

instead of :
bacpy(&loc_addr.l2_bdaddr, BDADDR_ANY);

Before I bind my address and PSM..

Thus binding my PSM and all to a hardcoded address instead of BDADDR_ANY..
My mistake! doh!..
Well sorry for all the fuss and mail spam I caused.. But thanks for all your
answers anyways. It is cool that a mailing list is functioning the way this
do: quick and good answers! :-)


/Steffen Larsen

----- Original Message -----
From: "Steffen Larsen" <zool@zool.dk>
To: <bluez-devel@lists.sourceforge.net>
Sent: Monday, December 13, 2004 10:48 AM
Subject: Re: [Bluez-devel] piconet..


>
> HI Marcel and other!,
>
> I found out what I did wrong. All my sockets and fork()'ing just worked
> fine. My problem was that I tried to make:
>
>
> ----- Original Message -----
> From: "Steffen Larsen" <zool@zool.dk>
> To: <bluez-devel@lists.sourceforge.net>
> Sent: Saturday, December 11, 2004 6:50 PM
> Subject: Re: [Bluez-devel] piconet..
>
>
> > Hi Marcel,
> >
> > > > I got the two others:
> > > >
> > > >  hci0: Type: USB
> > > > BD Address: 00:01:53:00:14:EE ACL MTU: 192:8 SCO MTU: 64:8
> > > > UP RUNNING PSCAN ISCAN
> > > > RX bytes:81 acl:0 sco:0 events:10 errors:0
> > > > TX bytes:39 acl:0 sco:0 commands:10 errors:0
> > > > Features: 0xff 0xff 0x0f 0x00 0x00 0x00 0x00 0x00
> > > > Packet type: DM1 DM3 DM5 DH1 DH3 DH5 HV1 HV2 HV3
> > > > Link policy:
> > > > Link mode: SLAVE ACCEPT
> > > > Name: 'knoglen'
> > > > Class: 0x000000
> > > > Service Classes: Unspecified
> > > > Device Class: Miscellaneous,
> > > > HCI Ver: 1.1 (0x1) HCI Rev: 0x20d LMP Ver: 1.1 (0x1) LMP Subver:
0x20d
> > > > Manufacturer: Cambridge Silicon Radio (10)
> > > >
> > > > hci0: Type: USB
> > > > BD Address: 00:01:53:00:0D:B5 ACL MTU: 192:8 SCO MTU: 64:8
> > > > UP RUNNING PSCAN ISCAN
> > > > RX bytes:87 acl:0 sco:0 events:11 errors:0
> > > > TX bytes:39 acl:0 sco:0 commands:10 errors:0
> > > > Features: 0xff 0xff 0x0f 0x00 0x00 0x00 0x00 0x00
> > > > Packet type: DM1 DM3 DM5 DH1 DH3 DH5 HV1 HV2 HV3
> > > > Link policy:
> > > > Link mode: SLAVE ACCEPT
> > > > Name: 'JESPER'
> > > > Class: 0x000000
> > > > Service Classes: Unspecified
> > > > Device Class: Miscellaneous,
> > > > HCI Ver: 1.1 (0x1) HCI Rev: 0x20d LMP Ver: 1.1 (0x1) LMP Subver:
0x20d
> > > > Manufacturer: Cambridge Silicon Radio (10)
> > >
> > > these two are HCI 16.4 and so there should be no problem. See my CSR
> > > page why it is good to have HCI 16.x at least.
> > >
> > > > > so far so good, but check with "hciconfig hci0 revision" for the
> > > > > firmware version. This one is too old. It should be HCI 16.4 or
> newer,
> > > > > but if this is a DBT-120 Rev. B3 you might have a chance to update
> it.
> > > >
> > > > Well. It is the B3 revision, so I would be able to do that. :-)
> > >
> > > I suppose you do that (at your own risk of course).
> > >
> > > > Hmm.. This dosn't really answer my question. I am using fork()., and
> > would
> > > > like to get my application to work. I tried to compare it towards
the
> > > > l2test, but couldn't see that much difference. And the l2test works
> with
> > the
> > > > same dongles; connecting two clients against one server. Can you
point
> > > > somewhere in my code, where I should do some changes, so it could
> work??
> > >
> > > Getting the fork stuff right is not very easy. You must carefully
think
> > > about what file descriptors to close at what time and in what process.
I
> > > am not going to check your code for such mistakes, because it will
take
> > > me too much time and the other reason is that it is C++ and not C.
> >
> > I understand. I have been spending alot of time in it my self. I will
look
> > deeper into the socket, and which of them that have been closed and at
> which
> > time... But I think I have checked this a million times.. But I will try
> > again! :-)
> >
> > > Update the dongle or try to swap the pre HCI 16.4 one. The master
option
> > > can help, but you must choose the master device carefully.
> >
> > Yes I can try to swap the devices. You mention that the master device
> should
> > be carefully choosen. I what way?
> > I have just assumed that my server should be the master, and thats that.
> > What should I taken in to consideration?
> >
> > -Once again thanks for all the quick answers! :-)
> >
> > /Steffen Larsen
> >
> >
> >
> > -------------------------------------------------------
> > SF email is sponsored by - The IT Product Guide
> > Read honest & candid reviews on hundreds of IT Products from real users.
> > Discover which products truly live up to the hype. Start reading now.
> > http://productguide.itmanagersjournal.com/
> > _______________________________________________
> > Bluez-devel mailing list
> > Bluez-devel@lists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/bluez-devel
> >
>
>
>
> -------------------------------------------------------
> SF email is sponsored by - The IT Product Guide
> Read honest & candid reviews on hundreds of IT Products from real users.
> Discover which products truly live up to the hype. Start reading now.
> http://productguide.itmanagersjournal.com/
> _______________________________________________
> Bluez-devel mailing list
> Bluez-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/bluez-devel
>



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/
_______________________________________________
Bluez-devel mailing list
Bluez-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bluez-devel

Re: [Bluez-devel] piconet..

[Marcel Holtmann]

Hi Steffen,

> My problem was that I used the:
> 
> bacpy(&loc_addr.l2_bdaddr, local_bdaddr);
> 
> instead of :
> bacpy(&loc_addr.l2_bdaddr, BDADDR_ANY);
> 
> Before I bind my address and PSM..
> 
> Thus binding my PSM and all to a hardcoded address instead of BDADDR_ANY..
> My mistake! doh!..

I don't see a problem here as long as you put in the local address of
your Bluetooth dongle (hci0) in there. If this is not working correctly
then we might have a kernel bug.

> Well sorry for all the fuss and mail spam I caused.. But thanks for all your
> answers anyways. It is cool that a mailing list is functioning the way this
> do: quick and good answers! :-)

You are welcome.

Regards

Marcel




-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/
_______________________________________________
Bluez-devel mailing list
Bluez-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bluez-devel