problem building kernel

This is merely a historical archive of years 2008-2021, before the migration to mailman3.

A maintained and still updated list archive can be found at https://lists.osmocom.org/hyperkitty/list/linux-dect@lists.osmocom.org/.

Romain Fliedel romain.fliedel at gmail.com
Fri Jul 29 16:54:30 UTC 2011


2011/7/29 Patrick McHardy <kaber at trash.net>

On 29.07.2011 15:27, Romain Fliedel wrote:
> 2011/7/28 Patrick McHardy <kaber at trash.net>
>
>> Am 28.07.2011 20:40, schrieb Romain Fliedel:
>>> 2011/7/28 Patrick McHardy <kaber at trash.net <mailto:kaber at trash.net>>
>>>
>>>     Am 28.07.2011 20:05, schrieb Romain Fliedel:
>>>     >
>>>     > 2011/7/28 Patrick McHardy <kaber at trash.net
>>>     <mailto:kaber at trash.net> <mailto:kaber at trash.net
>>>     <mailto:kaber at trash.net>>>
>>>     >
>>>     >     Am 28.07.2011 15:39, schrieb Fliedel Romain:
>>>     >     > Hi,
>>>     >     >
>>>     >     > I'm trying to build a kernel from the git repository with
>>>     dect stack
>>>     >     > enabled but i get this error :
>>>     >     >
>>>     >     > net/dect/ccp.c:25:27: fatal error: net/tipc/tipc.h:
>>>     >     >
>>>     >     > it appears that this file has been removed in kernel 2.6.36
>>>     >     >
>>>     >     > Do you have any idea to solve this problem ?
>>>     >     >
>>>     >
>>>     >     Disable the TIPC option for now, the code hasn't been adapter
>> to
>>>     >     the latest upstream changes yet.
>>>     >
>>>     >
>>>     > Thanks, that's what I end up doing and I am now able to build a
>>>     kernel.
>>>     > But now I have a problem with the driver, in my dmesg I've got the
>>>     > following error :
>>>     >
>>>     > com_on_air_pci: probe of 0000:07:04.0 failed with error -2
>>>     >
>>>     > any idea ?
>>>
>>>     I guess you linked the driver statically. This currently doesn't
>>>     work since the transceiver layer hasn't initialized the radio band
>>>     frequencies when the driver loads. Will probably fix that shortly,
>>>     using a modular build should work.
>>>
>>>
>>> Ok I'll try with the modular version.
>>> I noticed that no /dev/coa is created, is it normal ?
>>
>> Yes, the stack uses netlink and a socket API for userspace communication.
>>
>>
> Ok with the modular build it is working, thanks for you help.
> Now I'm trying to get a dump of the messages exchanged between a my FP and
a
> my PP, i've found the dectmon utility, but the dump file created is not
> usable in wireshark. Is there a way to convert it to a pcap file ?

No, but it should be easy to add. Alternatively you can use the
patched libpcap with tcpdump or wireshark directly. IIRC it needs
a one-liner change to the wireshark build though, will try to
find the patch again :)

 I've build wireshark with the modified libpcap, but I can't find a dect
interface. (even though I can produce a dump with dectmon)

If i call the function pcap_findalldevs() it return :

Devive : eth0
Desc : (null)
Devive : eth1
Desc : (null)
Devive : usbmon1
Desc : USB bus number 1
Devive : usbmon2
Desc : USB bus number 2
Devive : usbmon3
Desc : USB bus number 3
Devive : usbmon4
Desc : USB bus number 4
Devive : usbmon5
Desc : USB bus number 5
Devive : usbmon6
Desc : USB bus number 6
Devive : usbmon7
Desc : USB bus number 7
Devive : usbmon8
Desc : USB bus number 8
Devive : any
Desc : Pseudo-device that captures on all interfaces
Devive : lo
Desc : (null)

am I missing something ?

My configuration is the one provided on the wiki :

./dect-cluster-add --name cluster0 --mode fp --emc 0x1182 --fpn 0x0fac3
./dect-cell-add --name cell0 --cluster cluster0
./dect-transceiver-bind --transceiver trx0 --cell cell0
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osmocom.org/pipermail/linux-dect/attachments/20110729/0b91d9e6/attachment.htm>


More information about the linux-dect mailing list