**** BEGIN LOGGING AT Mon Jun 20 02:59:57 2011
Jun 20 10:00:07 <qlaude> the g
Jun 20 10:00:56 <qlaude> the g_idle_add functions which are extensively used while performing SIM IO is not wrking properly for me
Jun 20 10:01:08 <qlaude> any suggestions
Jun 20 10:18:40 <Rishi__> What is the role of Modem Emulator ??
Jun 20 13:56:42 <qlaude> hi guys!! i am facing a little problem. the functions called by g_idle_add are being called with some delay while performing simio.I have no clue ?Any help???
Jun 20 13:57:55 <denkenz> they are supposed to be called with some delay?
Jun 20 14:00:52 <qlaude> but they are called after like 5 mins delay.Is that possible
Jun 20 14:02:51 <qlaude> but when i am running it on my PC with phonesim it is not taking any delay
Jun 20 14:03:08 <denkenz> how old is your modem, are you querying the network in the meantime?
Jun 20 14:05:18 <qlaude> the network notifications are regularly coming but I am not handling them in ofono
Jun 20 14:05:27 <qlaude> driver right now
Jun 20 14:06:34 <denkenz> I mean are you querying any settings at the network, like CLIR/CLIP/CCWA/CCFC
Jun 20 14:07:00 <denkenz> And older modems take ages to read from the SIM, since oFono reads quite a few SIM files things do take a while
Jun 20 14:07:00 <qlaude> no
Jun 20 14:07:22 <qlaude> no i am getting the response from the sim right away
Jun 20 14:07:27 <denkenz> Though I've never seen it take 5 minutes
Jun 20 14:07:59 <qlaude> its only in ofono core that i am facing the problem
Jun 20 14:08:17 <denkenz> doubtful if you have it working properly in phonesim
Jun 20 14:08:42 <qlaude> the function scheduled to be called sim_fs_op_next is getting called after some delay
Jun 20 14:08:59 <qlaude> so where could the problem lie
Jun 20 14:10:05 <denkenz> can't really tell without a timestamped AT trace
Jun 20 14:11:06 <qlaude> the response from the modem is already decoded by the driver code but then it doesnot proceed further
Jun 20 14:12:14 <qlaude> like after getting the info of a specified file it takes ages to read the file
Jun 20 14:12:43 <denkenz> Reading it from where? The disk cache or the SIM?
Jun 20 14:13:00 <qlaude> from the sim
Jun 20 14:13:23 <denkenz> Then what does oFono have to do with it?
Jun 20 14:14:11 <qlaude> i am asking why are the g_idle_add funcs not called
Jun 20 14:14:49 <denkenz> They are on every single hw I have tried ;)
Jun 20 14:15:00 <denkenz> I suspect your modem is just too slow
Jun 20 14:15:22 <denkenz> I can't really help you more without a detailed AT trace
Jun 20 14:15:23 <qlaude> but i am getting the response from the modem
Jun 20 14:15:57 <qlaude> one more query why g_timeout_add funcs were not used
Jun 20 14:16:09 <qlaude> if i may ask
Jun 20 14:16:11 <qlaude> :)
Jun 20 14:16:14 <denkenz> where?
Jun 20 14:16:26 <qlaude> in place of g_idle_add
Jun 20 14:17:31 <denkenz> g_idle_add has the same effect as setting a 0 timeout
Jun 20 14:17:38 <denkenz> so what exactly are you asking?
Jun 20 14:18:30 <qlaude> ok sorry i thought g_idle_timeout funcs had better clarity
Jun 20 14:18:38 <qlaude> control*
Jun 20 14:19:19 <denkenz> If you're not doing anything evil with the event loop in your modem driver, then g_idle_add is just as good
Jun 20 14:20:10 <qlaude> like what?
Jun 20 14:20:37 <qlaude> the driver code has not much to do with the event loop i suppose
Jun 20 14:21:11 <denkenz> I dunno, reading a 2 GB file in a timeout or something :P
Jun 20 14:21:38 <denkenz> seriously, sim_fs* is not your problem, especially if it works on phonesim
Jun 20 14:21:49 <qlaude> no man!! but as you said i will check my driver code and then get back to you
Jun 20 14:22:01 <qlaude> thanks for your time :)
Jun 20 14:26:40 <qlaude> hey one query can i check the no of pending events in the event loop ??
Jun 20 14:30:04 <denkenz> no idea
Jun 20 14:30:40 <qlaude> neway thanks for ur time
Jun 20 15:37:26 <holtmann> denkenz: Actually g_idle_add and g_timeout_add are different.
Jun 20 15:37:53 <holtmann> g_idle_add will only be executed if no other IO functions are present.
Jun 20 15:40:50 <denkenz> That makes no sense
Jun 20 15:45:27 <holtmann> It makes sense, but in normal situation it makes no difference since IO is not under high load.
Jun 20 15:46:01 <denkenz> No, the idle events are just lower priority, not that they're not executed
Jun 20 15:47:22 <holtmann> They will be executed, but it can be delayed to last.
Jun 20 15:47:29 <holtmann> Except you keep IO busy.
Jun 20 15:47:37 <holtmann> timeouts should fire properly on time.
Jun 20 15:48:14 <denkenz> If the GLIB event loop is like every other event loop, then I/O will simply delay the execution of idle events in that iteration of the loop
Jun 20 15:48:38 <holtmann> That is what I am saying. That is why a timeout(0) is not the same as idle().
Jun 20 15:50:47 <denkenz> Not necessarily, I'm pretty sure both g_io_channel_add_watch and g_timeout_add uses PRIORITY_DEFAULT
Jun 20 15:50:55 <denkenz> so its just luck what executes first
Jun 20 15:51:09 <denkenz> unless glib uses some randomization internally
Jun 20 16:02:58 <holtmann> GLib itself prioritizes timeouts over idles.
Jun 20 16:03:56 <denkenz> sure, but using idle_add vs timeout_add doesn't really guarantee anything
**** ENDING LOGGING AT Tue Jun 21 02:59:56 2011