23 Jun
2010
23 Jun
'10
10:09 a.m.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hi, List!
It seems not possible to send SMS anymore with the yesterday's version
of OpenBSC (0.9.0.889-f7a1c).
I tryed with another old version (0.9.0.531-b938) and it works (same
NanoBTS, same configuration).
If I activate the log on Telnet I get:
<0002> gsm_04_08.c:621 <- CM SERVICE REQUEST serv_type=0x04
mi_type=0x04 M(1552623826) <0012> db.c:159 DBI: 1: ambiguous column
name: updated <0002> gsm_04_08.c:572 -> CM SERVICE ACK
<0012> db.c:159 DBI: 1: ambiguous column name: updated
<0004> abis_rsl.c:1313 RF release on (bts=0,trx=0,ts=2,ss=0) but state
ACTIVE <0004> abis_rsl.c:1024 (bts=0,trx=0,ts=2,ss=0) CHAN REL ACK but
state ACTIVE <0004> abis_rsl.c:1313 RF release on
(bts=0,trx=0,ts=2,ss=0) but state NONE <0004> abis_rsl.c:1024
(bts=0,trx=0,ts=2,ss=0) CHAN REL ACK but state NONE
on the old version, and:
<0002> gsm_04_08.c:770 <- CM SERVICE REQUEST serv_type=0x04
mi_type=0x04 M(1910630704) <0002> gsm_04_08.c:695 -> CM SERVICE ACK
<0004> abis_rsl.c:1350 RF release on (bts=0,trx=0,ts=2,ss=0) but state
ACTIVE <0004> abis_rsl.c:1039 (bts=0,trx=0,ts=2,ss=0) CHAN REL ACK but
state ACTIVE <0004> abis_rsl.c:1350 RF release on
(bts=0,trx=0,ts=2,ss=0) but state NONE <0004> abis_rsl.c:1039
(bts=0,trx=0,ts=2,ss=0) CHAN REL ACK but state NONE
on the new one.
I see, the line 572 (or 695) in gsm_04_08.c, and it is the same
function (gsm48_tx_mm_serv_ack), but with other parameters.
Has someone else the same problem? How can I solve it?
Thanks a lot
- --
_______________________________________________________________________
Luca Bertoncello
Entwicklung Mail: bertoncello(a)netzing.de
NETZING Solutions AG Tel.: 0351/41381 - 0
Kesselsdorfer Str. 216, 01169 Dresden Fax: 0351/41381 - 12
HRB 18926 / Ust.ID DE211326547 Mail: netzing.ag(a)netzing.de
_______________________________________________________________________
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
iD8DBQFMIbMRAXzltVKV/2wRAufHAJ9YkFNYhEXzdNSCuNwDEJcdoICWaACg5SD8
vITTtkdXrlqbxQOHFnoBpeE=
=69Za
-----END PGP SIGNATURE-----
23 Jun
23 Jun
10:18 a.m.
On 06/23/2010 03:09 PM, Luca Bertoncello wrote:
<0002> gsm_04_08.c:621 <- CM SERVICE REQUEST
serv_type=0x04
mi_type=0x04 M(1552623826) <0012> db.c:159 DBI: 1: ambiguous column
name: updated
is this noise? at least this looks like a difference to the old version...
Has someone else the same problem? How can I solve it?
You could try git bisect and see which commit broke that, but I think
one should understand the "new" DBI error first?
Could you also give us the "old" and "new" git hashes so we can take
a
look at what changed?
10:35 a.m.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Am Wed, 23 Jun 2010 15:18:41 +0800
schrieb Holger Freyther <zecke(a)selfish.org>rg>:
I think it, too! See my E-Mails about this problem... :D
On 06/23/2010 03:09 PM, Luca Bertoncello wrote:
Do you mean "ambiguous column name: updated"? This is not noise, this
is another bug. See my E-Mail on 18.06.2010 and then my solving
proposal on 22.06.2010.
The "new" version uses my patch (I'll send it to the list, if I'm sure
it works). I tried without my patch, too (last version from git,
0.9.0.890-2788), and it does not work, too.
<0002> gsm_04_08.c:621 <- CM SERVICE
REQUEST serv_type=0x04
mi_type=0x04 M(1552623826) <0012> db.c:159 DBI: 1: ambiguous column
name: updated
is this noise? at least this looks like a difference to the old
version...
Has someone
else the same problem? How can I solve it?
You could try git bisect and see which commit broke that, but I think
one should understand the "new" DBI error first?
Could you also give us the "old" and
"new" git hashes so we can take a
look at what changed?
Of course, but you have to say me how can I do it... I don't know git
so good... :(
Thanks
- --
_______________________________________________________________________
Luca Bertoncello
Entwicklung Mail: bertoncello(a)netzing.de
NETZING Solutions AG Tel.: 0351/41381 - 0
Kesselsdorfer Str. 216, 01169 Dresden Fax: 0351/41381 - 12
HRB 18926 / Ust.ID DE211326547 Mail: netzing.ag(a)netzing.de
_______________________________________________________________________
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
iD8DBQFMIblUAXzltVKV/2wRAogCAJ9jofBq0qhmmG9ENqd/QT0CdCuRugCg0qxL
hzn8q9k9Tcmp3R9Y9e6EJWQ=
=CDti
-----END PGP SIGNATURE-----
11:40 a.m.
On 06/23/2010 03:35 PM, Luca Bertoncello wrote:
Hi Luca,
I do not have a lot of spare time today. So let us please start over and
collect the information you have right now.
What are you doing?
Please provide instructions that can be easily repeated. This should
include if your phone did a successful Location Update (LU) and then
what you actually tried to do.
What do you expect?
Well, we kind of know that. You try to send a SMS from the phone and
then... well nothing...
What result did you get?
Well, SMS sending from the phone does not work.
And then please make sure to provide the full log (e.g. start bsc_hack
with -e1 and attach that in both cases) and also record a trace with
tcpdump of successful and unsuccessful sending...
PS:I just submitted a SMS from my N900 to an imaginary subscriber, I
have the log attached...
12:08 p.m.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Am Wed, 23 Jun 2010 16:40:35 +0800
schrieb Holger Hans Peter Freyther <holger(a)freyther.de>de>:
I do not have a lot of spare time today. So let us
please start over
and collect the information you have right now.
What are you doing?
Please provide instructions that can be easily repeated. This should
include if your phone did a successful Location Update (LU) and then
what you actually tried to do.
I just start the last version of OpenBSC. Two mobiles phone are logged
on my network. I can call them and speak, without any problem.
What do you expect?
Well, we kind of know that. You try to send a SMS from the phone and
then... well nothing...
Exactly! I send an SMS from a mobile to the other. It takes a very long
time, then said me, that it was impossible to send the SMS.
In the DB (table SMS) I find at least 3 SMS, with the content I tried
to send.
What result did you get?
Well, SMS sending from the phone does not work.
Exact! :D
And then please make sure to provide the full log
(e.g. start bsc_hack
with -e1 and attach that in both cases) and also record a trace with
tcpdump of successful and unsuccessful sending...
Here as Attachment the Logs of OpenBSC and TCPDump (zipped because of
the size).
Thanks a lot!
- --
_______________________________________________________________________
Luca Bertoncello
Entwicklung Mail: bertoncello(a)netzing.de
NETZING Solutions AG Tel.: 0351/41381 - 0
Kesselsdorfer Str. 216, 01169 Dresden Fax: 0351/41381 - 12
HRB 18926 / Ust.ID DE211326547 Mail: netzing.ag(a)netzing.de
_______________________________________________________________________
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
iD8DBQFMIc8gAXzltVKV/2wRAoNkAJ9W8g7HNrjm6yjTzM/i79uDJXJ6mwCcDlop
iCBALNMrVV0XuW/5iFpqALA=
=RoF5
-----END PGP SIGNATURE-----
12:59 p.m.
On 06/23/2010 05:08 PM, Luca Bertoncello wrote:
Exact! :D
The easiest is to use git bisect.
git bisect start -- openbsc (to limit it to that subdir)
git bisect good GOOD_REVISION (the sha1 not the stuff before the hyphen)
git bisect bad BAD_REVISION..
then repeat...
$ make
$ test_it...
and then either
$ git bisect good or git bisect bad...
at the end git will show you the commit that introduced the problem...
3:10 p.m.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Am Wed, 23 Jun 2010 17:59:47 +0800
schrieb Holger Hans Peter Freyther <holger(a)freyther.de>de>:
On 06/23/2010 05:08 PM, Luca Bertoncello wrote:
Hi,
I don't know git. Could you please help me and say how can I get the two
revision hashes? I just know the BSC version numbers.
Thanks
- --
_______________________________________________________________________
Luca Bertoncello
Entwicklung Mail: bertoncello(a)netzing.de
NETZING Solutions AG Tel.: 0351/41381 - 0
Kesselsdorfer Str. 216, 01169 Dresden Fax: 0351/41381 - 12
HRB 18926 / Ust.ID DE211326547 Mail: netzing.ag(a)netzing.de
_______________________________________________________________________
Exact! :D
The easiest is to use git bisect.
git bisect start -- openbsc (to limit it to that subdir)
git bisect good GOOD_REVISION (the sha1 not the stuff before the
hyphen) git bisect bad BAD_REVISION..
then repeat...
$ make
$ test_it...
and then either
$ git bisect good or git bisect bad...
at the end git will show you the commit that introduced the problem...
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
iD8DBQFMIfnDAXzltVKV/2wRAvXXAJ9QU3QCHkQyzlX1++n6/uA5Z1QEOACgwyZ0
sHU52E6cIua/69dIU+JsCxY=
=hgNM
-----END PGP SIGNATURE-----
10:53 a.m.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Am Wed, 23 Jun 2010 09:46:32 +0200
schrieb Sylvain Munaut <246tnt(a)gmail.com>om>:
Just because, in the new version, I applied a patch from mine, to
correct this error (just added a table name to the column names)...
Aha! Then I sent it already! The old version, the "sms sending
version", is 531-b938, and the new, the "non sms sending version",
889-f7a1c.
Thanks
- --
_______________________________________________________________________
Luca Bertoncello
Entwicklung Mail: bertoncello(a)netzing.de
NETZING Solutions AG Tel.: 0351/41381 - 0
Kesselsdorfer Str. 216, 01169 Dresden Fax: 0351/41381 - 12
HRB 18926 / Ust.ID DE211326547 Mail: netzing.ag(a)netzing.de
_______________________________________________________________________
You could try
git bisect and see which commit broke that, but I
think one should understand the "new" DBI error first?
From the original message I thought those error were in the old one
but not the new.
But in any case, the OP can fix the sql. The OP seem
to have a
new/more picky version of sqlite. Before in case of ambiguous column
name, it took the one from the first table in the FROM list. (not sql
standard but that's why it didn't matter before).
I know, it is NOT SQL-standard. And this is the reason, because I get
this error.
But I think, the program should use syntactically correct SQL-queries,
to avoid these problems.
I'll send my patch in a separated E-Mail (I hope today).
Could you also
give us the "old" and "new" git hashes so we can
take a look at what changed?
The hash (short) are in the version string, the log is :
git log b938..f7a1c
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
iD8DBQFMIb1oAXzltVKV/2wRAhwhAKCoSFzEi6ZP+QisWP1rce4u5k3VvACg5nnb
ajh0ciRccoEJi7HSqzEUa0E=
=R+OJ
-----END PGP SIGNATURE-----
5093
days inactive
5093
days old
9 comments
5 participants
participants (5)
-
Holger Freyther -
Holger Hans Peter Freyther -
Luca Bertoncello -
Nordin -
Sylvain Munaut