**** BEGIN LOGGING AT Mon Aug 02 02:59:57 2010
Aug 02 16:29:45 <mikeleib> tmzt: meegotouchframework, aava mobile
Aug 02 17:11:14 <padovan> denkenz: when you say re-use the atmodem GPRS stuff in dunmodem, what does that mean?
Aug 02 17:11:50 <denkenz> padovan: The ppp context setup for AT modems should be almost the same for DUN
Aug 02 17:12:48 <denkenz> In theory you can just add the atmodem gprs-context to the dunmodem gprs atom and have things just work
Aug 02 17:22:13 <padovan> Nice, I'll try that.
Aug 02 17:26:23 <denkenz> padovan: The DUN modem patches look like a decent start, but you have to check with holtmann whether he really wants you to use Serial proxy
Aug 02 17:26:43 <denkenz> It might be the use of RFCOMM sockets directly might be better
Aug 02 17:28:07 <padovan> Yeah, we can change that later, it's not critical to have DUN working.
Aug 02 17:28:36 <padovan> We can discuss this on the meeting as well.
Aug 02 17:32:15 <denkenz> fair enough
Aug 02 17:32:36 <denkenz> otherwise, try adding a basic gprs, netreg atoms and re-using the gprs-context from at modem
Aug 02 17:35:12 <denkenz> If it works, then we can start getting this stuff upstream
Aug 02 18:43:51 <holtmann> denkenz: I think that Serial Proxy is a nice start. We will figure out the best way to get around it soon. Mainly avoid libbluetooth is the key part.
Aug 02 18:44:03 <holtmann> denkenz: However if we do that, then in a way that it also helps obexd.
Aug 02 18:57:43 <denkenz> holtmann: The issue here is that there's conflicting feedback going to zhenhua and padovan
Aug 02 18:58:01 <denkenz> holtmann: In the end both are needing the same thing, namely RFCOMM sockets from btio
Aug 02 18:58:20 <denkenz> libbluetooth is not really necessary except for bdaddr2str and such, which can be easily copied
Aug 02 19:00:00 <holtmann> We will discuss this at the Bluetooth mini-summit.
Aug 02 19:00:32 <holtmann> Problem is getting this coordinated between the trees. I have to sit down and actually think about it a bit. And I need Johan's feedback on this.
Aug 02 19:01:01 <denkenz> Fair enough, but you should really say so to zhenhua / mailing list
Aug 02 19:02:06 <holtmann> He was discussing this with Johan on #bluez yesterday.
Aug 02 19:02:23 <holtmann> And I mentioned that Johan and I will discuss this end of the week.
Aug 02 19:02:24 <denkenz> Who deferred to you ;)
Aug 02 19:04:19 <holtmann> <holtmann> We need to talk about this next week. I first have to look a bit into the details on how we could achieve this.
Aug 02 19:04:58 <holtmann> So main idea is to make btio.[ch] Independent from libbluetooth and then copy it around. So we can avoid some nasty dependencies.
Aug 02 19:05:09 <holtmann> In fact in doing so it becomes useful for obexd as well.
Aug 02 19:05:25 <holtmann> Only problem is that it will carry L2CAP code which is not needed for oFono, but that is it.
Aug 02 19:05:46 <holtmann> Other part would be to split it in btio_rfcomm, btio_l2cap and then do compiler magic.
Aug 02 19:05:57 <holtmann> I haven't made up my mind yet.
Aug 02 19:06:58 <denkenz> No worries, my backlog must have missed your reply on #bluez and I haven't seen anything on the ML
Aug 02 19:07:39 <holtmann> I only skimmed the BlueZ mailing list for kernel patches for the 2.6.36 merge window. That has a bit higher priority right now.
Aug 02 19:08:13 <holtmann> I am bit busy right now :(
Aug 02 19:08:44 <denkenz> Nod, no worries
Aug 02 19:09:37 <holtmann> sameo: This reminds me. You do have a f35xx card?
Aug 02 19:10:00 <holtmann> Or just the f36xx one from Denis?
Aug 02 19:10:47 <holtmann> The DHCP library fails on the f35xx cards. Martin is looking into it, but I like to get this fixed before the next release. Their internal DHCP server must be a bit stupid.
Aug 02 19:11:00 <holtmann> sameo: So if you could have a second look why, that would help.
Aug 02 19:11:46 <holtmann> denkenz: You ever that including a DHCP server in 3G firmware was good idea must have smoked something really good ;)
Aug 02 19:12:00 <holtmann> s/You ever/Who ever thought/
Aug 02 19:12:17 <denkenz> There
Aug 02 19:12:24 <denkenz> 's no standard way of getting the IP/DNS
Aug 02 19:12:31 <denkenz> So in some way it makes sense
Aug 02 19:13:27 <holtmann> Someone really needs to sit down and either define a binary protocol or extend GSM 07.07 with the real life needs.
Aug 02 19:14:18 <denkenz> Its called ISI
Aug 02 19:15:21 <denkenz> Thought that has a few skeletons in the closet as well
Aug 02 19:15:28 <denkenz> s/thought/though
Aug 02 19:17:38 <holtmann> No being an official public standard and only one piece of hardware to begin with ;)
Aug 02 19:17:44 <holtmann> Anyhow, different topic.
Aug 02 19:17:52 <holtmann> oFono release, what is pending?
Aug 02 19:18:06 <denkenz> Nada
Aug 02 19:18:17 <holtmann> So I should just go ahead and do it?
Aug 02 19:18:31 <denkenz> Yeah, sounds like a plan
Aug 02 19:18:41 <denkenz> We can claim PoC STK support ;)
Aug 02 19:21:58 <holtmann> I tried it and had the wrong SIM card in my device. One that has no menus :(
Aug 02 19:22:04 <holtmann> And I was too lazy to swap it out ;)
Aug 02 19:22:28 <holtmann> I am telling you, I see more and more operators moving away from SIM menus.
Aug 02 19:23:08 <lizardo> holtmann: not in brazil, I'm afraid :(
Aug 02 19:23:10 <denkenz> Not the established ones
Aug 02 19:23:36 <denkenz> And some features are stupid in 3GPP, like MSP
Aug 02 19:23:46 <denkenz> Which is why Orange uses STK for their implementation
Aug 02 19:24:07 <holtmann> Rogers in Canada is done with them. So I wouldn't be surprised if more of them follow.
Aug 02 19:24:10 <denkenz> There are other examples too, like all the crappy STK stuff Australian operators do
Aug 02 19:24:29 <holtmann> Until all of them give up, we have to do it anyway, but we are working on it. So be it.
Aug 02 19:24:45 <denkenz> Yeah, and some operators never had it, like AT&T
Aug 02 19:25:13 <denkenz> Btw, you can test with phonesim
Aug 02 19:25:21 <denkenz> It has a fully enabled test app
Aug 02 19:26:02 <holtmann> Nice. I saw that you pushed one extra patch into phonesim. Do we need more. Or should I just release this along with oFono to have a combo that works with each other?
Aug 02 19:26:05 <denkenz> Just hit 'Start' in the SIM tab
Aug 02 19:26:18 <denkenz> The current combination is enough
Aug 02 19:26:32 <holtmann> commit 79626031b3551626dbcbd3d7e6820985bebaa077
Aug 02 19:26:32 <holtmann> Author: Denis Kenzior <denkenz@gmail.com>
Aug 02 19:26:32 <holtmann> Date:   Fri Jul 30 17:10:46 2010 -0500
Aug 02 19:26:32 <holtmann>     Add end session support
Aug 02 19:26:35 <denkenz> However, once we add GetInput, GetInkey and a few other commands more will be required in Phonesim
Aug 02 19:26:37 <holtmann> This is not needed?
Aug 02 19:26:42 <denkenz> No, that one is needed
Aug 02 19:26:49 <holtmann> Okay. Then we wait.
Aug 02 19:27:02 <denkenz> Basically upstream oFono & upstream phonesim work
Aug 02 19:27:23 <denkenz> Once more things are added to oFono, things might need to be tweaked in phonesim
Aug 02 19:27:35 <denkenz> But that'll take some time
Aug 02 19:28:17 <denkenz> Since its 2 weeks since last release, might be fine to release both phonesim and ofono
Aug 02 19:29:47 <holtmann> Phonesim difference is only the end session patch. I did 1.3 phonesim without telling anybody ;)
Aug 02 19:30:37 <denkenz> Yeah I know, but phonesim is basically 'done'
Aug 02 19:30:45 <denkenz> So 1 patch is fine for a release
Aug 02 19:30:48 <holtmann> The lets do a 1.4 as well.
Aug 02 19:31:02 <denkenz> Anyway, up to you, but definitely do an oFono release so QA can have an early look
Aug 02 19:31:21 <holtmann> Pre-release test builds are already running.
Aug 02 19:31:21 <kristenc>               M                                                                                                                                                        
Aug 02 19:31:44 <kristenc> denkenz: do you have any suggestions for the best way to test sim operations?
Aug 02 19:32:16 <denkenz> kristenc: Phonesim
Aug 02 19:32:39 <denkenz> kristenc: Pretty sure that one handles offset-ed reads
Aug 02 19:32:58 <denkenz> kristenc: Might need to double check whether it handles 256 byte reads properly
Aug 02 19:33:19 <kristenc> denkenz: thanks.  I've forgotten - is there documentation somewhere on how to setup phonesim?
Aug 02 19:33:35 <denkenz> see phonesim/HACKING
Aug 02 19:33:50 <kristenc> ok
Aug 02 19:35:11 <denkenz> kristenc: Just paste this into modem.conf once you have phonesim running and start ofono:
Aug 02 19:35:12 <denkenz> [phonesim]
Aug 02 19:35:12 <denkenz> Driver=phonesim
Aug 02 19:35:12 <denkenz> Address=127.0.0.1
Aug 02 19:35:13 <denkenz> Port=12345
Aug 02 19:35:29 <kristenc> ok
Aug 02 20:08:27 <jprvita> denkenz: busy? do you have time to have a look on the voiceonly-sim patch I just wrote: http://pastebin.com/UnGmd8Mg ?
Aug 02 20:19:24 <kristenc> hum, so I am not doing something right.  I put modem.conf into /etc/ofono, and when I run test/list-modems I get:
Aug 02 20:19:28 <kristenc> [ /phonesim ]
Aug 02 20:19:28 <kristenc>     Interfaces = 
Aug 02 20:19:28 <kristenc>     Powered = 0
Aug 02 20:19:28 <kristenc>     Features = 
Aug 02 20:19:28 <kristenc>     Online = 0
Aug 02 20:19:49 <kristenc> so if I type ./enable-modem /phonesim I get errors.  What am I missing?
Aug 02 20:23:34 <denkenz> Did you start phonesim?
Aug 02 20:36:06 <kristenc> denkenz: no - I thought I had to run this script first.
Aug 02 20:36:25 <denkenz> oFono connects to phonesim when you enable the modem
Aug 02 20:36:43 <denkenz> so phonesim must be running before you run the script
Aug 02 20:36:52 <sameo> holtmann: I'll try to look
Aug 02 20:37:32 <kristenc> denkenz: thanks, that seems to be working.
Aug 02 20:37:35 <holtmann> sameo: Thanks.
Aug 02 20:38:41 <jprvita> denkenz: which other sim state values can I expect from huawei sysinfo? right now I'm only aware of 1, 3, 4 and 255
Aug 02 20:38:52 <jprvita> and what's the exact meaning of 1?
Aug 02 20:38:58 <holtmann> martin_lab: ofono-0.26 and phonesim-1.4 are uploading right now.
Aug 02 20:39:44 <denkenz> jprvita: Oh sorry, forgot to look at the patch, let me do it now
Aug 02 20:39:52 <denkenz> jprvita: There are I think 0-4 and 255
Aug 02 20:40:01 <jprvita> no worries :)
Aug 02 20:40:11 <denkenz> Gimme a sec to pull up the docs
Aug 02 20:40:28 <jprvita> I didn't do the "query sysinfo a few times" part yet
Aug 02 20:42:07 <denkenz> jprvita: Yep, I see it
Aug 02 20:42:48 <denkenz> Looking good to me, though the part in notify_sim_state is a bit wrong
Aug 02 20:43:06 <denkenz> Let me paste the meanings of sim state to see why
Aug 02 20:44:40 <jprvita> ok
Aug 02 20:44:42 <kristenc> denkenz: is something supposed to happen when I hit the "start" button on the SIM tab?
Aug 02 20:45:06 <denkenz> kristenc: Yeah, that's for SIM Toolkit
Aug 02 20:45:21 <denkenz> kristenc: But only needed if you're playing with the new Sim Toolkit api
Aug 02 20:45:30 <denkenz> jprvita: So sim_state: 0, 1, 2, 3, 4, 255
Aug 02 20:45:39 <denkenz> 0 -> Invalid USIM Card state or PIN Code locked
Aug 02 20:45:43 <denkenz> 1 -> Valid USIM card state
Aug 02 20:45:51 <denkenz> 2 -> USIM is invalid in case of CS
Aug 02 20:45:57 <denkenz> 3 -> USIM is invalid in case of PS
Aug 02 20:46:03 <denkenz> 4 -> USIM is invalid in case of CS or PS
Aug 02 20:46:09 <denkenz> 255 -> USIM card is not existent
Aug 02 20:47:01 <jprvita> so, PS == data and CS == voice?
Aug 02 20:47:06 <denkenz> Correct
Aug 02 20:47:46 <denkenz> So the funny part is that you can have a PIN locked card
Aug 02 20:47:53 <denkenz> so it would start in mode 0
Aug 02 20:47:57 <denkenz> And then could go into mode 4
Aug 02 20:48:21 <jprvita> shouldn't we notify sim as inserted for all except 255 and check for other issues the same way I did with gprs?
Aug 02 20:48:23 <denkenz> The current driver is handling this wrongly on many levels
Aug 02 20:48:35 <denkenz> Yep, that is what I'm thinking as well
Aug 02 20:50:31 <jprvita> for this plugin, features related with CS are only SMS and CBS? (although I don't know what is this last one)
Aug 02 20:51:06 <denkenz> No, CS would only be voice
Aug 02 20:51:18 <denkenz> So essentially the same ones enabled for EM770
Aug 02 20:52:03 <denkenz> However, for the case of sim_state == 4, I suggest disabling everything except sim and phonebook
Aug 02 20:52:14 <jprvita> everything under "
Aug 02 20:52:27 <jprvita> if (ofono_modem_get_boolean(modem, "HasVoice") == TRUE)
Aug 02 20:52:34 <denkenz> yep
Aug 02 20:52:34 <jprvita> (CS stuff)
Aug 02 20:52:49 <denkenz> In the case of 4, you basically can't do anything with the SIM except maybe poke the SIM card
Aug 02 20:53:50 <jprvita> and this 'HasVoice' is a device property then?
Aug 02 20:53:57 <jprvita> not sim property
Aug 02 20:54:26 <denkenz> Correct, its enabled for devices we know support voice
Aug 02 20:54:44 <denkenz> as in actual support, not the you can make voice calls but never hear the sound
Aug 02 20:56:58 <jprvita> in that case what is missing so we can actually hear the sound?
Aug 02 20:57:49 <jprvita> (if I got your sentence correctly)
Aug 02 21:02:36 <denkenz> The hookup to the sound chip
Aug 02 21:02:48 <denkenz> Most USB devices simply never hook this up
Aug 02 21:03:00 <denkenz> You'd need a separate serial channel, like N900 does
Aug 02 21:05:57 <denkenz> Ok, I'm off for a bit
Aug 02 21:06:07 <jprvita> ok, thanks for the explanations
Aug 02 21:06:16 <jprvita> tomorrow I'll send you a new patch
Aug 02 21:29:56 <inakypg> denkenz: ping
Aug 02 21:31:56 <inakypg> denkenz: ping
Aug 02 22:00:16 <holtmann> inakypg: I would assume it is dinner time in Texas ;)
Aug 02 22:20:36 <inakypg> Marcel: weird early dinners!
Aug 02 22:46:40 <sameo> holtmann: I only have an f36xx
Aug 02 22:47:46 <holtmann> sameo: Then you have to hack oFono to also use DHCP for f36xx. It is possible.
Aug 02 22:48:00 <holtmann> We just don't wanna do that by default.
Aug 02 22:49:40 <sameo> holtmann: If you have an ofono patch handy, we'd save some time.
Aug 02 22:52:44 <sameo> would hacking settings->static_ip be enough ?
Aug 02 22:55:38 <sameo> doesnt seem so...
Aug 02 22:59:43 <holtmann> In drivers/mbmmodem/gprs-context.c
Aug 02 23:00:03 <holtmann> mbm_e2ipcfg_query_cb() change  gcd->have_e2ipcfg = FALSE.
Aug 02 23:03:32 <sameo> ok, I was there...
Aug 02 23:04:24 <holtmann> oFono is actually smart and checks the support for static ip configuration.
Aug 02 23:04:43 <holtmann> So you have to manually overwrite it to tell it that the hardware doesn't support it.
Aug 02 23:05:20 <holtmann> If the f35xx would support static IP configuration, we would use it there as well. It is just that only the f36xx and later does :(
Aug 02 23:05:26 <luke-jr> is Nokia back now⁇
Aug 02 23:06:57 <sameo> holtmann: gcd->have_e2ipcfg = FALSE doesnt seem to work
Aug 02 23:08:18 <holtmann> diff --git a/drivers/mbmmodem/gprs-context.c b/drivers/mbmmodem/gprs-context.c
Aug 02 23:08:19 <holtmann> index cca5087..e08156f 100644
Aug 02 23:08:19 <holtmann> --- a/drivers/mbmmodem/gprs-context.c
Aug 02 23:08:19 <holtmann> +++ b/drivers/mbmmodem/gprs-context.c
Aug 02 23:08:21 <holtmann> @@ -434,7 +434,7 @@ static void mbm_e2ipcfg_query_cb(gboolean ok, GAtResult *res
Aug 02 23:08:21 <holtmann>         struct ofono_gprs_context *gc = user_data;
Aug 02 23:08:23 <holtmann>         struct gprs_context_data *gcd = ofono_gprs_context_get_data(gc);
Aug 02 23:08:24 <holtmann>  
Aug 02 23:08:26 <holtmann> -       gcd->have_e2ipcfg = ok;
Aug 02 23:08:27 <holtmann> +       gcd->have_e2ipcfg = FALSE;
Aug 02 23:08:29 <holtmann>  }
Aug 02 23:08:59 <sameo> exactly
Aug 02 23:09:17 <holtmann> That would be a major bug.
Aug 02 23:09:25 <holtmann> make clean please and rebuild ofonod.
Aug 02 23:10:34 <holtmann> Can you run with OFONO_AT_DEBUG=1 and paste the trace of AT commands.
Aug 02 23:10:46 <holtmann> Maybe there is a race condition in the code somewhere.
Aug 02 23:10:50 <sameo> sure
Aug 02 23:11:00 <holtmann> I don't have f36xx with me. So I can't test this.
Aug 02 23:16:35 <sameo> holtmann: http://pastebin.com/EBH65rQF
Aug 02 23:19:03 <holtmann> This looks like it does the right thing.
Aug 02 23:19:16 <holtmann> Run monitor-ofono to see its PropertyChanged D-Bus messages.
Aug 02 23:19:44 <sameo> holtmann: and I correctly get dhcp as the method from the PrimaryDataContext interface
Aug 02 23:20:22 <holtmann> So it should be fine from oFono point of view now.
Aug 02 23:20:30 <holtmann> Just that our dhcp-lib doesn't work.
Aug 02 23:20:59 <sameo> let me track that down
Aug 02 23:21:23 <holtmann> Btw, there is one other bug on ConnMan here. Even if oFono requires us to use DHCP, the IPv4.Method = fixed should show.
Aug 02 23:21:50 <holtmann> Since the end user can't just overwrite the 3G IP settings. They are given by the network.
Aug 03 00:01:23 <sameo> holtmann: this is quite fucked. We do receive a DHCP offer, but then our DHCP request is never answered
Aug 03 00:02:02 <sameo> holtmann: dhclient receives the same DHCP offer and replies with a packet wireshark knows nothing about, and then magically gets a DHCP ACK
Aug 03 00:04:39 <luke-jr> lol
Aug 03 00:04:47 <sameo> dhclient sends a DHCP request, but the protocol bits are set to 0, as opposed to 0x0800 for dhcp-lib
Aug 03 00:08:50 <luke-jr> seriously, when is Nokia coming back? :/
Aug 03 00:35:27 <sameo> holtmann: dhcp issue fixed
Aug 03 01:53:51 * luke-jr grumbles.
**** ENDING LOGGING AT Tue Aug 03 02:59:56 2010