**** BEGIN LOGGING AT Wed Nov 25 03:35:38 2009 Nov 25 04:23:01 hi guys Nov 25 06:23:39 denkenz: Linux ubuntu 2.6.31-14-generic Nov 25 06:25:33 Sage: Strange, I'm running 2.6.31 on Gentoo Nov 25 06:25:45 Sage: Can you give me any more information, perhaps step through in debugger? Nov 25 06:26:00 Sage: Make sure to re-bootstrap using CFLAGS=-O0 Nov 25 06:27:43 Sage: Since the Option 401 card 'just' works for me, but maybe the udev setup isn't correct for your card or something Nov 25 06:31:25 denkenz: does ofono autodetect the modem on your system? Nov 25 06:31:52 I undertood that it should, but at least on my system it does not. Nov 25 06:32:14 It does, but ofono.rules for udev might not be correct Nov 25 06:32:48 Sage: AFAIK right now we support the 225 and the 401 Nov 25 06:33:03 I don't have the other cards Nov 25 06:33:16 Can you edit plugins/ofono.rules to add your productid and restart udev? Nov 25 06:33:43 Copy the rules file to the relevant directory too before restarting Nov 25 06:34:43 denkenz: one question Nov 25 06:34:49 zhenhua1: yep? Nov 25 06:34:53 I'll see what I can do. I didn't know that I need to copy that ofono.rules file to the udev/rules.d :) Nov 25 06:35:02 denkenz: in manager_release_and_answer and hold_and_answer Nov 25 06:35:15 denkenz: why we check waiting call Nov 25 06:35:29 Sage: Yeah, that's what contains all the udev magic that tells oFono what modem driver to use Nov 25 06:36:20 zhenhua1: Mostly in the case of having a held + active + waiting call Nov 25 06:36:28 zhenhua1: In that case you can't use hold + answer Nov 25 06:36:56 denkenz: i know, but you may have held+active. then you're wrong Nov 25 06:37:14 denkenz: in 3-way call, swap two calls between held and active Nov 25 06:37:18 zhenhua1: In which case you should just accept :) Nov 25 06:37:49 denkenz: accept what? PTS has such test cases.... Nov 25 06:38:16 denkenz: waiting -> active ->hold Nov 25 06:38:37 denkenz: using CHLD=1 or 2 Nov 25 06:39:44 zhenhua1: Ok, lets do this one at a time Nov 25 06:39:46 denkenz: shall we change if (!voicecalls_have_waiting(vc)) to if (voicecalls_have_waiting(vc)) Nov 25 06:39:57 zhenhua1: manager_hold_and_answer Nov 25 06:40:02 yeah Nov 25 06:40:19 If we don't have a waiting call here, we fail Nov 25 06:40:25 The application should be using SwapCalls Nov 25 06:40:34 If we have held + active + waiting, we fail Nov 25 06:40:40 Since that is not a valid situation Nov 25 06:41:00 Make sense? Nov 25 06:41:07 yes, i know Nov 25 06:41:15 then what about release_and_answer Nov 25 06:41:39 btw, i think we can merge swapcalls and holdandanswer Nov 25 06:42:02 So in this case if we don't have waiting calls we should just be using Hangup or HangupMultiparty Nov 25 06:43:02 i know. then i choose the wrong method to invoke Nov 25 06:43:19 Remember, we're trying to make it make sense for the user Nov 25 06:43:39 The way GSM setup the CHLD invocation is bizarre at best Nov 25 06:43:46 completely utterly stupid at worst Nov 25 06:43:53 heh Nov 25 06:44:09 yes. i am confusing it with CHLD=1, 2 Nov 25 06:45:07 So you see the logic now? That's about the best way I could come up with that made sense to a non-gsm geek Nov 25 06:47:34 yes Nov 25 06:48:08 denkenz: found the problem. Nov 25 06:48:29 Sage: What was it? Nov 25 06:48:38 denkenz: If you do not have udev development files the ofono compiles but does not work :) Nov 25 06:48:42 checking for UDEV... no Nov 25 06:49:14 Sage: Ah, that would do it Nov 25 06:49:43 That is so small print during the configuration that I haven't noticed before. Nov 25 06:50:05 Sage: Nod, I got bit by that in the beginning, so no worries Nov 25 06:50:28 Btw, should the devices work with the modem.conf even when the udev isn't included to build? Nov 25 06:50:28 Sage: If you require a new entry to ofono.rules, send a patch please with a comment about the card that you have Nov 25 06:50:42 Sage: No, modem.conf is a bit different Nov 25 06:51:03 Sage: udev is for these USB & such devices which are pre-multiplexed Nov 25 06:51:15 Sage: modem.conf is mostly for old serial devices which need a muxer Nov 25 06:51:18 denkenz: Then, I think that the compile should fail if udev development file are not found? Nov 25 06:51:38 Sage: You can still use it without udev, so I don't necessarily agree Nov 25 06:51:46 denkenz: I see. Nov 25 06:52:42 holtmann: Maybe we should include --enable-udev in bootstrap-configure? Nov 25 06:52:51 holtmann: That might alleviate some of the pain Nov 25 06:56:23 holtmann, denkenz: I think that --enable-udev is a good idea, because then users would notice it better. Nov 25 06:57:00 holtmann, denkenz: Btw, there seems to be --disable-udev already, so should udev be enabled by default? Nov 25 06:57:49 Sage: If it detects udev, it will enable it by default Nov 25 06:58:12 Sage: The behavior of configure is actually correct the way it is now Nov 25 06:58:36 Sage: For developers it might be better to include --enable-udev in bootstrap-configure Nov 25 07:01:45 for running on a device I need a static option for configuring ports Nov 25 07:02:00 static option? Nov 25 07:03:22 like modems.conf Nov 25 07:03:54 We have to see, modem.conf is about as ugly as it can be already Nov 25 07:05:02 I just mean there's no vendprod for instance Nov 25 07:05:25 vendprod? Nov 25 07:05:30 if udev is used a script would still be needed to register /dev/smd7 as at port Nov 25 07:05:40 like usb uses, vend:prod Nov 25 07:05:55 Ah, yeah the MSM devices are interesting Nov 25 07:06:01 Do they actually have multiple ports? Nov 25 07:06:02 not a script, a udev or hal fdi or whatever Nov 25 07:06:07 yes Nov 25 07:06:20 Are they always the same name or different names? Nov 25 07:06:23 the windows based ones need a port for ppp Nov 25 07:06:39 they should always be the same name Nov 25 07:06:53 In that case I suggest to simply write a custom driver for msm ones Nov 25 07:08:06 You can take Device to mean a base name, then append the port id however you wish Nov 25 07:09:29 I see Nov 25 07:09:43 I should be working on cdma/gsm touch pro 2 soon Nov 25 07:09:49 kernel work first though Nov 25 07:39:41 denkenz: ofono.rules patch for couple of option modems available at the ML. Nov 25 08:00:21 ping denkenz Nov 25 08:01:38 denkenz: i am still thinking to change release_and_answer. Nov 25 08:02:06 denkenz: if we have hold+active call, now we cannot use CHLD=1 to release active. but it's possible. Nov 25 10:10:05 zhenhua: Sure, its possible, but what's the point when you can use hangup? Nov 25 10:13:06 ath? Nov 25 10:13:28 it drops data connection on some modems Nov 25 10:13:56 zhenhua1: I am going to introduce hangup (only active calls) and hangup_all (hangup active + held, but not waiting) Nov 25 10:15:54 tmzt: Yeah I know about the ath stuff interfering with GPRS and data calls on some devices Nov 25 10:22:20 zhenhua1: If we were to do the active + held case, we'd make a new method called ReleaseAndSwap, but I seriously question its usefulness right now Nov 25 23:22:02 holtmann: So I have a Virgin and 3 prepaid SIMs Nov 25 23:22:10 holtmann: Let me know what you wanna know Nov 26 00:18:46 denkenz: why you need new ReleaseAndSwap, simply remove waiting call check in ReleaseAndAnswer should be fine from my point. Nov 26 00:20:15 That's because you know what CHLD=1 does Nov 26 00:21:44 Anyway, I'm not convinced right now, unless someone can make a compelling usecase **** ENDING LOGGING AT Thu Nov 26 02:59:56 2009