**** BEGIN LOGGING AT Thu Oct 21 02:59:58 2010
Oct 21 04:41:49 <matt_chen> hi guys, how would I use ofono to enable a 3G module ?
Oct 21 04:42:10 <matt_chen> it always shows me "No services found"
Oct 21 06:01:13 <akiniemi> pessi: just use the ISI version to quirk the N95 support
Oct 21 06:06:29 <matt_chen> hi, how would I make sure my modem is enabled ?
Oct 21 06:06:55 <matt_chen> I am using MeeGo
Oct 21 06:07:31 <matt_chen> with ofono v0.20-1.2
Oct 21 06:08:02 <matt_chen> or any tutorial to test the modem ?
Oct 21 08:15:37 <balrog-k1n> matt_chen: try downloading the source for that version and then running the test/list-modems script is the easiest way to read all the modem properties ofono knows about
Oct 21 08:18:43 <matt_chen> balrog-k1n : I've done that and also saw the modem info.
Oct 21 08:19:04 <matt_chen> balrog-k1n : next is how to dial a number and get an IP from ISP ?
Oct 21 08:19:39 <matt_chen> balrog-k1n : needs to set the username and password as well
Oct 21 08:21:28 <balrog-k1n> matt_chen: there should be another script in test/ to do that, but maybe someone else here knows the details
Oct 21 08:22:00 <matt_chen> balrog-k1n : yes, thanks ...
Oct 21 08:22:21 <balrog-k1n> there were many improvements in that area since ofono v20 though
Oct 21 08:23:04 <matt_chen> suggest updated ofono ?
Oct 21 08:24:35 <balrog-k1n> yes, but if you want to use it together with the other packages in meego, like connman and the GUIs (i don't really know which those are) you also need to update thoso to compatible versions
Oct 21 08:24:47 <balrog-k1n> not if you just want to test ofono
Oct 21 08:25:32 <matt_chen> hmm...I didn't figure it is that complicated though....no wonder see on UI in carrick.
Oct 21 18:32:59 <akiniemi> Oops. gatchat/gatchat.c:1317: error: ‘return’ with no value, in function returning non-void
Oct 21 18:37:13 <holtmann> akiniemi: My bad.
Oct 21 18:37:16 <holtmann> I fix it.
Oct 21 18:38:16 <holtmann> Pushed.
Oct 21 18:42:39 <akiniemi> Thanks
Oct 21 20:47:58 <holtmann> akiniemi: Do you have pending patches?
Oct 21 21:07:41 <akiniemi> holtmann: nope, nothing ready
Oct 21 21:08:01 <holtmann> I am considering cutting another release.
Oct 21 21:08:06 <akiniemi> holtmann: but Pekka's set of SIM patches sure would be nice, as they include isimodem PIN support
Oct 21 21:08:52 <holtmann> I think they are still in review. I wait for Denis when he is back online and see how far he is with the review.
Oct 21 21:09:27 <akiniemi> ok
Oct 21 21:33:10 <holtmann> denkenz: So do we have pending patches? I am thinking of another release.
Oct 21 21:36:05 <denkenz> holtmann: there are some things I was planning to integrate today
Oct 21 21:36:16 <denkenz> But they're not particularly critical
Oct 21 21:36:34 <holtmann> I can do tomorrow as well.
Oct 21 21:36:40 <holtmann> Which ones do you wanna work on?
Oct 21 21:36:55 <denkenz> I was going to merge the rest of Yang's patches and possibly Petteri's
Oct 21 21:37:06 <holtmann> Okay. Then do that.
Oct 21 21:37:13 <holtmann> Then we have a nice release.
Oct 21 21:38:29 <holtmann> denkenz: Another thing is how do I enable multiple context setup?
Oct 21 21:39:32 <denkenz>                         if (ctx->context.cid !=
Oct 21 21:39:32 <denkenz>                                         idmap_get_min(ctx->gprs->cid_map)) {
Oct 21 21:39:32 <denkenz>                                 ofono_error("Multiple active contexts are"
Oct 21 21:39:33 <denkenz>                                                 " not yet supported");
Oct 21 21:39:36 <denkenz> Take out that if
Oct 21 21:40:08 <holtmann> I have. Breaks.
Oct 21 21:40:24 <holtmann> Second call to active second context stalls.
Oct 21 21:41:00 <holtmann> Do I have to add a second GPRS context driver to it as well? I have done that since we need another chat instance.
Oct 21 21:41:32 <denkenz> yes you do
Oct 21 21:41:45 <holtmann> I have done that. Doesn't work.
Oct 21 21:42:45 <denkenz> Dunno then
Oct 21 21:43:20 <denkenz> I'm pretty sure it is as simple as adding two contexts and taking out that if
Oct 21 21:44:36 <denkenz> But then since I never had hw to test it in the first place, maybe not all the plumbing is completely there
Oct 21 21:45:32 <holtmann> Maybe with ISI it is, but for me it seems we are using the wrong contexts.
Oct 21 21:47:05 <holtmann> ofonod[29342]: GPRS2: > AT+CGDATA="M-RAW_IP",1\r
Oct 21 21:47:05 <holtmann> ofonod[29342]: GPRS2: < \r\nCONNECT\r\n
Oct 21 21:47:36 <holtmann> And then I get this:
Oct 21 21:47:38 <holtmann> [root@localhost ~]# ./test/deactivate-context
Oct 21 21:47:38 <holtmann> Error activating /ifx0/context1: org.ofono.Error.InProgress: Operation already in progress
Oct 21 21:47:56 <holtmann> So first it uses the second GPRS context first.
Oct 21 21:48:01 <holtmann> Then deactivating it fails.
Oct 21 21:48:42 <denkenz> then probably the plumbing needs more work
Oct 21 21:49:02 <denkenz> maybe we don't actually select between the context drivers
Oct 21 21:50:04 <holtmann> And in case of ISI it just works because they do things differently.
Oct 21 21:52:47 <denkenz> maybe
Oct 21 21:52:58 <denkenz> I suspect we don't maintain the list of context drivers
Oct 21 21:53:04 <holtmann> Can you have a look at that.
Oct 21 21:53:06 <denkenz> and just use the last one
Oct 21 21:53:31 <denkenz> sure, its in my queue
Oct 21 21:54:24 <holtmann> It could be also that I did something wrong in the driver.
Oct 21 21:54:44 <holtmann> But I don't see any AT commands coming on that second channel. So I assume not.
Oct 21 21:54:46 <denkenz> doesn't look like it
Oct 21 22:04:51 <holtmann> denkenz: So the code is all present. Just src/gprs.c needs to be fixed.
Oct 21 22:05:04 <holtmann> And maybe ifxmodem/gprs-context.c a bit.
Oct 21 22:14:31 <holtmann> struct ofono_gprs_context *context_driver;
Oct 21 22:14:38 <holtmann> It is not a list, it is just one driver.
Oct 21 22:14:53 <holtmann> So the second one overwrites the first one.
Oct 21 22:18:12 <raji__> denkenz, just fyi, that the latest ofono code doesnt compile due to warnings in maintainer mode.
Oct 21 22:18:51 <holtmann> raji__: I know. Pushed a fix already. The MeeGo compiler flags for building 32-bit binaries disabled my sanity check.
Oct 21 22:18:54 <holtmann> All blame on me.
Oct 21 22:23:14 <denkenz> holtmann:
Oct 21 22:23:17 <denkenz>         gcd = g_new0(struct gprs_context_data, 1);
Oct 21 22:23:17 <denkenz> +       if (!gcd)
Oct 21 22:23:17 <denkenz> +               return -ENOMEM;
Oct 21 22:23:18 <denkenz> +
Oct 21 22:23:23 <denkenz> forgot to use g_try_new0
Oct 21 22:23:43 <holtmann> Crap. Second time today. Will fix.
Oct 21 22:24:26 <denkenz> And I still don't like the set_slave bit
Oct 21 22:25:22 <holtmann> I like it a lot actually. I played with all ideas left and right. This is simple and nice.
Oct 21 22:26:04 <denkenz> I would have preferred group, slave feels like a hack
Oct 21 22:27:24 <holtmann> As I said, I had the GAtChatPair first.
Oct 21 22:27:48 <holtmann> Didn't like it.
Oct 21 22:28:57 <denkenz> Then you might wanna add docs to it
Oct 21 22:29:12 <denkenz> Personally it doesn't pass my 'its obvious from the function name' test
Oct 21 22:29:29 <holtmann> Fair enough. I can do that.
Oct 21 22:35:43 <holtmann> denkenz: Just send you a crash that I had. For some reason it doesn't make it to the mailing list :(
Oct 21 22:37:21 <denkenz> that crash makes no sense
Oct 21 22:37:26 <denkenz> I need the rest of the log
Oct 21 22:38:06 <holtmann> Where should I get that one from :(
Oct 21 22:38:46 <holtmann>   CC     gisi/client.o
Oct 21 22:38:46 <holtmann> ../gisi/client.c:38:17: error: log.h: No such file or directory
Oct 21 22:38:46 <holtmann> ../gisi/client.c: In function ‘build_subscribe_msg’:
Oct 21 22:38:46 <holtmann> ../gisi/client.c:614: warning: implicit declaration of function ‘DBG’
Oct 21 22:39:01 <holtmann> akiniemi: Please always run also fakeroot make distcheck.
Oct 21 22:39:53 <denkenz> your crash dump must be from a non-debug build
Oct 21 22:40:09 <holtmann> Took it on the maschine.
Oct 21 22:42:11 <denkenz> The only reason you should see dbus complain is if the mcc/mnc of the operator was screwed up
Oct 21 22:42:20 <holtmann> akiniemi: I removed the DBG calls and the inclusion of log.h. That is not allowed in gisi/
Oct 21 22:42:31 <holtmann> If you want the debug, then you have to do that somehow else.
Oct 21 22:42:36 <denkenz> see if you can duplicate it and give me the rest of the AT log
Oct 21 22:42:52 <holtmann> Lets try. Just doing sanity build checks.
Oct 21 22:46:13 <holtmann> ofonod[7788]: GPRS2: > AT+CGDATA="M-RAW_IP",1\r
Oct 21 22:46:13 <holtmann> ofonod[7788]: GPRS2: < \r\nCONNECT\r\n
Oct 21 22:46:13 <holtmann> ofonod[7788]: Net: < \r\n+XCIEV: 3,\r\n
Oct 21 22:46:13 <holtmann> ofonod[7788]: Net: < \r\n+XCIEV: 2,\r\n
Oct 21 22:46:15 <holtmann> ofonod[7788]: Net: < \r\n+XCIEV: 1,\r\n
Oct 21 22:46:15 <holtmann> ofonod[7788]: Net: < \r\n+XCIEV: 0,\r\n
Oct 21 22:46:17 <holtmann> ofonod[7788]: Net: < \r\n+CREG: 2,"560e","6e32",3\r\n
Oct 21 22:46:17 <holtmann> ofonod[7788]: Net: > AT+CGATT=0\r
Oct 21 22:46:19 <holtmann> ofonod[7788]: Net: < \r\n+CGREG: 2\r\n
Oct 21 22:46:22 <holtmann> Now it is fine :(
Oct 21 22:46:51 <denkenz> somehow the valid_path dbus check got triggered
Oct 21 22:47:04 <denkenz> The only way that's feasible is if the mcc/mnc somehow had garbage in it
Oct 21 22:48:41 <holtmann> Maybe I was in between a cell switch.
Oct 21 22:49:03 <holtmann> ofonod[24512]: Net: > AT+COPS=?\r
Oct 21 22:49:03 <holtmann> ofonod[24512]: Net: < \r\n+COPS: (2,"T-Mobile D","TMO D","26201",0,0),(3,"E-Plus
Oct 21 22:49:03 <holtmann> ","E-Plus","26203",0,),(3,"Vodafone.de","Vodafone.de","26202",0,),(3,"o2 - de","
Oct 21 22:49:04 <holtmann> o2 - de","26207",0,)\r\n
Oct 21 22:49:06 <holtmann> ofonod[24512]: Net: < \r\nOK\r\n
Oct 21 22:49:08 <holtmann> Btw. What is wrong with this one?
Oct 21 22:49:19 <holtmann> scan-for-operators does not work.
Oct 21 22:50:25 <denkenz> If you have the original log with the crash, see what COPS? returned
Oct 21 22:51:06 <holtmann> So this is funny. If it sees T-Mobile then the operator list fails:
Oct 21 22:51:07 <holtmann> ofonod[7788]: Net: < \r\n+COPS: (2,"T-Mobile D","TMO D","26201",0,0),(3,"Vodafone.de","Vodafone.de","26202",0,),(3,"o2 - de","o2 - de","26207",0,),(3,"E-Plus","E-Plus","26203",0,)\r\n
Oct 21 22:51:25 <holtmann> If not, then it is fine:
Oct 21 22:51:26 <holtmann> ofonod[7788]: Net: < \r\n+COPS: (2,"E-Plus","E-Plus","26203",0,),(3,"o2 - de","o2 - de","26207",0,),(3,"Vodafone.de","Vodafone.de","26202",0,)\r\n
Oct 21 22:52:14 <denkenz> T-mobile is reporting double 0
Oct 21 22:52:24 <denkenz> We only expect a single number after the mccmnc
Oct 21 22:53:18 <holtmann> The others have the comma. What is the expected value there. And why do all have a comma.
Oct 21 22:53:35 <denkenz> 27.007 is fuzzy here, but in general each tech should be on a separate line
Oct 21 22:53:50 <denkenz> +COPS: [list of supported (<stat>,long alphanumeric <oper>,short alphanumeric <oper>,numeric <oper>[,<AcT>])s][,,(list of supported <mode>s),(list of supported <format>s)]
Oct 21 22:54:26 <denkenz> The 's after AcT is refering to operators, not AcTs
Oct 21 22:54:42 <denkenz> seems IFX is interpreting the spec a bit differently
Oct 21 22:55:00 <denkenz> Or at least that's the way I read it
Oct 21 22:55:47 <holtmann> AT+COPS=? +COPS: [list of supported (<stat>,long alphanumeric <oper>
Oct 21 22:55:47 <holtmann>           ,short alphanumeric <oper>,numeric <oper>[,< AcT>,<plmn_list>)s]
Oct 21 22:55:47 <holtmann>           [,,(list of supported <mode>s),(list of supported <format>s)]
Oct 21 22:55:47 <holtmann>           OK
Oct 21 22:55:49 <holtmann>           or
Oct 21 22:55:49 <holtmann>           +CME ERROR: <err>
Oct 21 22:55:54 <holtmann> They have some plmn_list here.
Oct 21 22:56:13 <holtmann> T-Mobile is my home operator.
Oct 21 22:56:27 <denkenz> then you need to quirk IFX for that one
Oct 21 22:56:36 <denkenz> what's the meaning of plmn_list?
Oct 21 22:56:52 <holtmann> <plmn_list>
Oct 21 22:56:53 <holtmann> is supporting only in R7 Protocol Stack.
Oct 21 22:57:09 <holtmann> <plmn_list>
Oct 21 22:57:09 <holtmann> indicates whether the PLMN is present on the EHPLMN list, the User Controlled PLMN List or the Operator
Oct 21 22:57:09 <holtmann> Controlled PLMN List.
Oct 21 22:57:09 <holtmann> 0 PLMN is present on the EHPLMN list
Oct 21 22:57:10 <holtmann> 1 PLMN is present on the User Controlled PLMN List
Oct 21 22:57:10 <holtmann> 2 PLMN is present on the Operator Controlled PLMN List
Oct 21 22:57:47 <denkenz> so pointless and against the spec
Oct 21 23:02:07 <denkenz> akiniemi: Thanks for fixing my booboo btw, no idea how the compiler allowed that
Oct 21 23:03:38 <holtmann> denkenz: May a newer spec version?
Oct 21 23:06:38 <denkenz> Well, considering plmn_list talks about Rel7 features and the newest spec is Rel10...
Oct 21 23:06:42 <denkenz> But lets test your theory
Oct 21 23:08:26 <denkenz> holtmann: nope
Oct 21 23:08:28 <denkenz> Rel 10.1:
Oct 21 23:08:29 <denkenz> +COPS: [list of supported (<stat>,long alphanumeric <oper>,short alphanumeric <oper>,numeric <oper>[,<AcT>])s][,,(list of supported <mode>s),(list of supported <format>s)]
**** ENDING LOGGING AT Fri Oct 22 02:59:58 2010