23 Jun
2010
23 Jun
'10
12:13 p.m.
On 06/23/2010 06:01 PM, Luca Bertoncello wrote:
Am Wed, 23 Jun 2010 17:50:19 +0800
schrieb Holger Hans Peter Freyther <holger(a)freyther.de>de>:
if ((tar->filter_map & (1 << FLT_IMSI)) != 0 && subscr &&
strcmp(subscr->imsi, tar->filter_data[FLT_IMSI]) == 0)
return 1;
the above does not evaluate to true... this can be:
a) because the filter function is not called at all..
b) tar->filter_map & .. is == 0
c) subscr is NULL
d) the imsi's do not match..
Okay, after wanting to explain how the filtering is working, I
remembered that I did this: 69e8f8285bf080ad2050fbc20f861bc8621e5c75
(git show that-number). You could git revert that-number and then try again.
So in the case of the measurement report the subscr is not set as
context as the measurement debug is still in the abis_rsl context before
it is hitting the BSC side...
> Unfortunately it does not work...
> If I just give "logging filter imsi XXX" OpenBSC logs nothing.
> I have to give "logging filter all 1", too, and then both
> measurements will be sent.
Well,
look at src/debug.c and figure out why the imsi comparison is failing?
I'm looking now this code. But I can't understand the reason...
I can only suppose, that in this Measurement Report the IMSI will not
be sent.