**** BEGIN LOGGING AT Thu Feb 04 02:59:57 2010 Feb 04 04:38:30 denkenz: I saw you updated the overview.txt, we can add info about HFP there too Feb 04 04:46:59 padovan: i don't quite understand why you replace hfp_disable with set_powered(FALSE). Feb 04 04:47:24 padovan: it should be correct to call hfp_disable Feb 04 05:15:45 padovan: For once I agree with zhenhua here, you were mostly ok using hfp_disable Feb 04 05:16:59 padovan: the change to hfp_disable removing ofono_modem_set_powered is correct though Feb 04 05:17:47 Not sure how I missed that one :) Feb 04 05:21:09 However, using hfp_disable is ugly, I suggest a dedicated function for this Feb 04 05:21:21 So it is clear of the intent Feb 04 05:25:32 padovan: Ah I understand why you did this change, ok, so you need to do both Feb 04 05:26:08 padovan: Make a function that calls set_powered so core returns an error, AND also do the cleanup steps of hfp_disable Feb 04 05:27:05 calling set_powered is actually not strictly necessary, the core maintains its own timer and will report an error if it expires Feb 04 06:07:09 denkenz: does atserver need wakeup command? no, right? Feb 04 06:08:24 denkenz: it should be only used by client side, sending like 'AT' to server to wakeup server. Feb 04 06:20:11 zhenhua: no, the wakeup command thing is specific to TI Calypso, that one is just plain weird Feb 04 06:20:19 zhenhua: I know of no other modems it is required for Feb 04 06:20:28 zhenhua: Don't worry about it for server Feb 04 06:20:35 ok Feb 04 18:36:20 hi. what is the minimal glibc version for ofono? first i've tried it with 2.6.1, but no signalfd. now i'm trying it with 2.7 but some signalfd structs don't have the necessary members. any experiences? Feb 04 18:38:09 matgnt: Not really. Check man signalfd Feb 04 18:38:31 VERSIONS Feb 04 18:38:32 signalfd() is available on Linux since kernel 2.6.22. Working support Feb 04 18:38:32 is provided in glibc since version 2.8. The signalfd4() system call Feb 04 18:38:32 (see NOTES) is available on Linux since kernel 2.6.27. Feb 04 18:39:15 holtmann: ok. thanks Feb 04 18:52:13 padovan is generating more list traffic than everyone else combined :) Feb 04 18:53:42 denkenz: hahah. A lot of reviews on the code. :) Feb 04 19:19:40 padovan: I have the first set of cleanups sent to you, review them please Feb 04 19:22:13 denkenz: ok, I'll take a look Feb 04 19:26:24 padovan: The timeout handling was removed for now, we should instead set a proper timeout and callback on HandsfreeGateway.Connect Feb 04 19:28:12 Fine. Feb 04 19:32:25 +free: Feb 04 19:32:26 + if (data) { Feb 04 19:32:26 + g_free(data->handsfree_path); Feb 04 19:32:26 + g_free(data); Feb 04 19:32:26 + } Feb 04 19:32:45 denkenz: looks we don't need to free the handsfree_path here Feb 04 19:34:03 modem = ofono_modem_create(NULL, "hfp"); Feb 04 19:34:03 + if (modem == NULL) Feb 04 19:34:03 + goto free; Feb 04 19:34:25 denkenz: cant just return -ENOMEM here. Feb 04 19:34:48 and remove the if (modem) on the fre label Feb 04 19:36:16 denkenz: the service_level_conn_failed seems fine Feb 04 19:46:09 denkenz: http://pastebin.com/m73e83d5 some clean ups Feb 04 19:46:13 padovan: yeah agree with both Feb 04 19:50:51 padovan: Ok I pushed those up now Feb 04 19:51:55 denkenz: ok, I'll rebase here to get changes. Feb 04 19:52:20 padovan: Couple other things I noticed: Feb 04 19:52:44 hfp_disable always sends UnregisterAgent, even if potentially triggered from ReleaseAgent Feb 04 19:53:09 make that hfp_remove Feb 04 19:53:29 hfp_disable should call Disconnect and return -EINPROGRESS Feb 04 19:53:48 Once Disconnect succeeds it should call ofono_modem_set_powered Feb 04 19:54:55 Ok, I'll work on that. Feb 04 20:09:10 krau: Now lets start working on SAP, I do have a Telit UC864 board here Feb 04 20:33:35 denkenz, hi, I still waiting mine. It was ordered last week Feb 04 20:34:30 krau: no hurry, but you can start getting a SAP interface ready for BlueZ upstream Feb 04 20:34:37 would make my life easier Feb 04 20:35:21 denkenz, ok perfect. Feb 04 20:36:49 denkenz, what is the current status? Does oFono support it already? Feb 04 20:37:20 krau: I've gotten the Telit to run and its basically a compliant generic at modem Feb 04 20:37:43 krau: We mostly need to reuse the code from plugins/hfp and do some socket data forwarding Feb 04 20:38:12 I only did basic investigation so far though Feb 04 20:38:24 denkenz, telit has an office here is Brazil, so probably they will send me the EVK2 soon. Feb 04 20:39:32 fine, I will check plugins/hfp Feb 04 20:39:42 denkenz: who sends UnregisterAgent is hfp_remove. When we call Release it calls ofono_modem_remove() which calls hfp_remove. I think we can add a flag for registered status. Feb 04 20:39:44 Until then we can work on the plumbing for BlueZ side & oFono side Feb 04 20:40:04 padovan: Yeah, exactly what I had in mind Feb 04 20:40:13 =) Feb 04 20:41:12 padovan: You also need to do a g_dbus_unregister on the agent Feb 04 20:41:20 padovan: So that might be your 'flag' Feb 04 22:24:57 padovan: Also work on one that times out the connect call Feb 04 22:25:09 padovan: Set the timeout at 15 seconds or so Feb 04 22:25:37 padovan: And if connect fails, you probably want to send a Disconnect with a no-reply message Feb 04 22:33:23 denkenz: ok Feb 04 22:34:16 I think we'll be mostly OK after that, still curious why your pending call reference counting is different from what is in connman Feb 04 22:43:43 denkenz: What's the difference? Feb 04 22:45:53 padovan: You unref the pending call right away Feb 04 22:46:01 connman unrefs it when callback is received **** ENDING LOGGING AT Fri Feb 05 02:59:56 2010