gatttool and multiple devices connected

gatttool and multiple devices connected

[Kim Schulz]

Hi guys,
I was just testing the LE functionality a bit and decided to connect to 
two different LE devices at the same time using gatttool in interactive 
mode (two different terminals). Connection to the two devices went well, 
but when I tried to get primary services on the devices it was only 
possible to do so on one of the devices.
Connecting to them one by one showed that it was actually possible to 
get the primary service listing of both devices - just not when both are 
connected at the same time.
I guess this is a bug somewhere in the primary-service lookup code but 
have yet to look more into the details of this.

Just thought that you might wanted to know about this.


cheers
Kim

Re: gatttool and multiple devices connected

[Anderson Lizardo]

Hi Kim,

On Tue, Oct 16, 2012 at 11:28 AM, Kim Schulz <kim@schulz.dk> wrote:
> Hi guys,
> I was just testing the LE functionality a bit and decided to connect to two
> different LE devices at the same time using gatttool in interactive mode
> (two different terminals). Connection to the two devices went well, but when
> I tried to get primary services on the devices it was only possible to do so
> on one of the devices.
> Connecting to them one by one showed that it was actually possible to get
> the primary service listing of both devices - just not when both are
> connected at the same time.
> I guess this is a bug somewhere in the primary-service lookup code but have
> yet to look more into the details of this.
>
> Just thought that you might wanted to know about this.

Which kernel version are you running? Older kernels had a LE
connection routing issue which I believe was fixed long ago.

And what do you mean with not being able to get services? which error
did you get?

Do you have hcidump logs from these sessions?

Regards,
-- 
Anderson Lizardo
Instituto Nokia de Tecnologia - INdT
Manaus - Brazil

Re: gatttool and multiple devices connected

[Kim Schulz]

On 10/16/2012 04:45 PM, Anderson Lizardo wrote:
> Which kernel version are you running? Older kernels had a LE 
> connection routing issue which I believe was fixed long ago. And what 
> do you mean with not being able to get services? which error did you 
> get? Do you have hcidump logs from these sessions? Regards, 

The Bluez code is one taken from git a few days ago. I don't recall 
which kernel version (will have to check at the office tomorrow) but it 
is one of the latest that we compiled for this test.
We do have some logs of the session and I will be looking at those more 
thoroughly tomorrow.

kim

RE: gatttool and multiple devices connected

[Kim Schulz]

Den 2012-10-16 16:51, Kim Schulz skrev:
> On 10/16/2012 04:45 PM, Anderson Lizardo wrote:
>> Which kernel version are you running? Older kernels had a LE 
>> connection routing issue which I believe was fixed long ago. And what 
>> do you mean with not being able to get services? which error did you 
>> get? Do you have hcidump logs from these sessions? Regards,
>
> The Bluez code is one taken from git a few days ago. I don't recall 
> which kernel version (will have to check at the office tomorrow) but 
> it is one of the latest that we compiled for this test.
> We do have some logs of the session and I will be looking at those 
> more thoroughly tomorrow.
>
> kim
> --
> To unsubscribe from this list: send the line "unsubscribe 
> linux-bluetooth" in the body of a message to majordomo@vger.kernel.org 
> More majordomo info at  http://vger.kernel.org/majordomo-info.html


Hi,
Turns out that the guy compiling our kernel choose a version 
3.4.0.030400 so not really the
bleeding edge version.
Do you have any recommendation on what version of the kernel we should 
aim for?


-- 
Kim Schulz

Re: gatttool and multiple devices connected

[Anderson Lizardo]

Hi Kim,

On Wed, Oct 17, 2012 at 6:00 AM, Kim Schulz <kim@schulz.dk> wrote:
> Turns out that the guy compiling our kernel choose a version
> 3.4.0.030400 so not really the
> bleeding edge version.
> Do you have any recommendation on what version of the kernel we should
> aim for?

This kernel is actually not that old (as long as it is based on the
official 3.4.0 one). Did you have hcidump logs for the attempt to
discover primary services on device B while connected to device A?

Regards,
-- 
Anderson Lizardo
Instituto Nokia de Tecnologia - INdT
Manaus - Brazil

Re: gatttool and multiple devices connected

[Johan Hedberg]

Hi Lizardo,

On Wed, Oct 17, 2012, Anderson Lizardo wrote:
> On Wed, Oct 17, 2012 at 6:00 AM, Kim Schulz <kim@schulz.dk> wrote:
> > Turns out that the guy compiling our kernel choose a version
> > 3.4.0.030400 so not really the
> > bleeding edge version.
> > Do you have any recommendation on what version of the kernel we should
> > aim for?
> 
> This kernel is actually not that old (as long as it is based on the
> official 3.4.0 one). Did you have hcidump logs for the attempt to
> discover primary services on device B while connected to device A?

Actually I'd say that it is too old for LE usage. E.g. the address type
field to the L2CAP socket address was introduced in 3.5 which is quite
important from an LE standpoint.

Johan

Re: gatttool and multiple devices connected

[Anderson Lizardo]

Hi Johan,

On Wed, Oct 17, 2012 at 8:35 AM, Johan Hedberg <johan.hedberg@gmail.com> wrote:
>> This kernel is actually not that old (as long as it is based on the
>> official 3.4.0 one). Did you have hcidump logs for the attempt to
>> discover primary services on device B while connected to device A?
>
> Actually I'd say that it is too old for LE usage. E.g. the address type
> field to the L2CAP socket address was introduced in 3.5 which is quite
> important from an LE standpoint.

Ok, my mistake. I thought it was introduced still on 3.4.0. In this
case, Better use the latest mainline (3.6.*).

Regards,
-- 
Anderson Lizardo
Instituto Nokia de Tecnologia - INdT
Manaus - Brazil