| 00:27.07 | *** join/#asterisk-dev denon (i=denon@sassinak.net) |
| 00:34.06 | *** join/#asterisk-dev obnauticus (n=l@about/windows/regular/obnauticus) |
| 00:36.14 | moy | seanbright: so, I think the tapping stuff is stable and clean enough for review, should I open a bug report and create a review to propose it for inclusion? |
| 00:37.11 | JunK-Y | hey moy! |
| 00:37.14 | JunK-Y | tapping stuff? |
| 00:37.57 | moy | hey JunK-Y how ya doing? .... yeah, is some little feature I've been working on, it allows Asterisk to use physical taps in PRI lines to monitor calls |
| 00:38.59 | moy | http://svnview.digium.com/svn/asterisk/team/moy/dahdi-tap-trunk/ |
| 01:21.05 | *** join/#asterisk-dev Daviey (n=Daviey@ubuntu/member/pdpc.gold.Daviey) |
| 04:16.01 | *** join/#asterisk-dev denon (i=denon@sassinak.net) |
| 07:16.40 | *** join/#asterisk-dev obnauticus (n=obnautic@about/windows/regular/obnauticus) |
| 07:16.58 | *** join/#asterisk-dev tzafrir_laptop (n=tzafrir@local.xorcom.com) |
| 07:16.58 | *** mode/#asterisk-dev [+o tzafrir_laptop] by ChanServ |
| 08:15.17 | tzafrir_laptop | regarding my comments about tapping requires no hardware change: it turns out I remembered completely incorrectly |
| 08:15.23 | tzafrir_laptop | Ignore those |
| 08:17.38 | *** join/#asterisk-dev |Cybex| (n=John@80.100.126.176) |
| 09:04.15 | atis_home | could anybody explain why i see only few dozens of megabytes in "memory show summary", while asterisk is using gigabytes? http://pastebin.ca/1581183 |
| 09:36.37 | atis_home | is waiting for 5 gig core to get dumped |
| 11:27.26 | *** join/#asterisk-dev puzzled (n=foobar@puzzled.xs4all.nl) |
| 13:16.38 | *** join/#asterisk-dev oej (n=olle@ns.webway.se) |
| 13:16.38 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 13:46.00 | mvanbaak | wtf! 4.8GB memory usage for asterisk ? |
| 13:46.13 | mvanbaak | what are you doing with it man ? |
| 14:17.43 | *** join/#asterisk-dev tzafrir_laptop (n=tzafrir@212.179.75.202) |
| 14:17.43 | *** mode/#asterisk-dev [+o tzafrir_laptop] by ChanServ |
| 14:21.38 | Corydon76-dig | atis_home: yes. Due to Linux's memory management scheme, memory allocated to a process is never freed back to the general pool until the process ends. So your memory footprint represents the maximum amount of memory that your process has ever used. |
| 14:22.59 | Corydon76-dig | Additionally, if there's a memory leak somewhere (not always in Asterisk; sometimes it could be in a dependent library), that could cause malloc(3) to continually balloon the memory footprint |
| 14:38.06 | mvanbaak | Corydon76-dig: is that linux specific or is it also present in the BSD's ? |
| 15:39.55 | tzafrir_laptop | Though this is really a policy of glibc, not something set in stone in the kernel, right? |
| 16:10.46 | tzafrir_laptop | re-correct my correction: I see that the dahdi-tap code uses two spans for tapping. Which means one for each direction, I guess. |
| 16:11.03 | tzafrir_laptop | If that is the case, there are (almost) no hardware changes needed |
| 17:03.01 | *** join/#asterisk-dev leifmadsen (n=Leif@asterisk/documenteur-extraordinaire/blitzrage) |
| 17:03.01 | *** mode/#asterisk-dev [+o leifmadsen] by ChanServ |
| 17:29.45 | *** join/#asterisk-dev leifmadsen (n=Leif@asterisk/documenteur-extraordinaire/blitzrage) |
| 17:29.45 | *** mode/#asterisk-dev [+o leifmadsen] by ChanServ |
| 18:06.59 | atis_home | Corydon76-dig: well, if it's so, it's kind of stupid, because asterisk takes 9 gigs of ram, and can't do fork |
| 18:08.30 | tzafrir_laptop | atis_home, fork uses copy-on-write |
| 18:09.09 | tzafrir_laptop | if the fork is closely followed by an exec, the process will allocate pages from scratch |
| 18:09.55 | tzafrir_laptop | atis_home, also, in such a case, add a large swap file |
| 18:10.16 | tzafrir_laptop | most of those memory pages will be unused, and will happily reside in the swap |
| 18:10.28 | atis_home | well, i had System() dialplan app failing all inbound calls, because not enough memory |
| 18:11.52 | tzafrir_laptop | I'm also a bit surprised that the oom killer didn't kill asterisk |
| 18:11.58 | tzafrir_laptop | did you disable it? |
| 19:06.30 | *** join/#asterisk-dev leifmadsen (n=Leif@asterisk/documenteur-extraordinaire/blitzrage) |
| 19:06.30 | *** mode/#asterisk-dev [+o leifmadsen] by ChanServ |
| 19:18.13 | paravoid | Corydon76-dig: not exactly |
| 19:18.42 | paravoid | memory allocated is never freed back to other processes but, assuming there isn't a leak, it's available to subsequent malloc() calls |
| 19:23.11 | paravoid | so there shouldn't be a 5gb memory consumption unless that's what happened at some point |
| 19:24.09 | mvanbaak | snuff-home: ping |
| 19:49.32 | oej | mvanbaak: Still have compilation issues of 1.4 on FreeBSD... |
| 19:49.39 | oej | Doesn't find /usr/local/lib |
| 19:49.54 | mvanbaak | oej: that sux |
| 19:50.03 | oej | Yes, doesn't it... |
| 19:50.04 | mvanbaak | oej: I tested on OpenBSD 4.4, 4.5 and -current |
| 19:50.07 | mvanbaak | and there it works |
| 19:50.16 | oej | It used to work a few weeks ago, but stopped in an update |
| 19:50.30 | oej | I tested in FreeBSD 5 and 7 |
| 19:50.51 | mvanbaak | oej: you can compare the stuff for freebsd and openbsd |
| 19:51.01 | oej | I guess I have to... |
| 19:51.02 | mvanbaak | oej: I know there's some specific stuff for openbsd in the makefile |
| 19:51.03 | oej | Where to start? |
| 19:51.10 | oej | Ok |
| 19:51.37 | mvanbaak | oej: I cannot test on freebsd for some days |
| 19:51.47 | mvanbaak | oej: I have no fbsd virtualmachine anymore |
| 19:52.31 | oej | We set /usr/local/lib for all BSDs |
| 19:52.43 | oej | _ASTLDFLAGS |
| 19:52.53 | oej | Wonder why it doesn't work in FreeBSD but in OpenBSD |
| 19:54.27 | mvanbaak | because openbsd is superiour |
| 19:54.33 | oej | ha ha |
| 19:54.44 | mvanbaak | ;) |
| 20:07.11 | *** part/#asterisk-dev |Cybex| (n=John@80.100.126.176) |
| 20:07.29 | *** join/#asterisk-dev kpfleming (n=kpflemin@asterisk/digium-director-of-software-tech/kpfleming) |
| 20:07.29 | *** mode/#asterisk-dev [+o kpfleming] by ChanServ |
| 20:16.04 | *** join/#asterisk-dev tnt_ (n=tnt_@142.12-244-81.adsl-dyn.isp.belgacom.be) |
| 20:16.23 | tnt_ | Hi everyone. Is there a good tutorial on how to write a channel driver ? (that explains all those call backs :) |
| 20:25.01 | kpfleming | not really, no |
| 20:25.14 | kpfleming | we usually point people at chan_oss as an example, since it's a very simple channel driver |
| 20:25.54 | kpfleming | although it only supports a single active channel, so it's not an example of how to manage channel private structures in the driver itself |
| 20:33.51 | tnt_ | kpfleming: I've been looking at several of them but they seem to do things differently. For example in 'call', to tell asterisk it's ringing,some send a control frame, and some send a control frame _and_ call ast_indicate. Also, I'm not sure to understand what the 'fixup' call back is supposed to do. |
| 20:34.25 | kpfleming | sending the control frame is the right way to go |
| 20:34.34 | kpfleming | it will get turned into ast_indicate automatically if needed |
| 20:35.01 | kpfleming | fixup is used when a channel masquerade is done, and the 'guts' of one channel need to be moved into a different channel (usually after a transfer has occurred) |
| 20:35.36 | kpfleming | it's asterisk's way of informing the channel driver that a specific native channel (managed by the driver) now belongs to a different asterisk channel than it used to |
| 20:40.31 | tnt_ | kpfleming: Ok i see. I more question: Once a channel has been obtained by a 'requester' call or has been passed to asterisk via pbx_start(...), it will _always_ finish by a call to 'hangup' right (and that call will be the last) ? |
| 20:41.01 | kpfleming | yep |
| 20:41.38 | kpfleming | the driver's hangup callback will be called to hangup the channel and allow it to destroy its private data (or at least disconnect it from the asterisk channel, if it must hang around for a bit) |
| 20:45.47 | tnt_ | kpfleming: Ok great thanks. I think I have enough to go on and clean the messy hack I coded this afternoon :) |
| 20:46.18 | kpfleming | great |
| 20:46.25 | mvanbaak | wow, coding a complete channel driver in 1 afternoon |
| 20:47.35 | tnt_ | mvanbaak: Huh 'complete' might be a big word :) So far I can just make one call (and only one :) |
| 20:47.55 | tnt_ | but now I can clean what I have so far and go from there. |
| 20:48.24 | mvanbaak | gheh. |
| 20:48.29 | mvanbaak | good luck tnt_ ! |
| 20:49.40 | kpfleming | tnt_: actually, if you want another example, download the skypeforasterisk package. even if you don't use it, the source code for chan_skype is included, and it was written very recently so it should be a good example to follow, including support for multiple channels |
| 20:50.05 | kpfleming | you can't copy any of that code and use it, though, since it isn't GPL |
| 20:50.09 | mvanbaak | can you still download that ? |
| 20:50.24 | mvanbaak | last time I checked I had to pay for it |
| 20:51.16 | kpfleming | the package is downloadable from downloads.digium.com |
| 20:51.22 | kpfleming | you have to pay for licenses to use it |
| 20:51.31 | mvanbaak | echo "pretty please" | mail -s 'please move asterisk to the BSD license' legal@digium.com |
| 20:51.34 | mvanbaak | ;) |
| 20:51.59 | mvanbaak | ah. I only checked the website to see if I could find a download link |
| 20:52.18 | mvanbaak | did not look at downloads.digium.com |
| 20:52.33 | mvanbaak | my fault |
| 20:53.06 | mvanbaak | hhmm, SP2 for windows 2003 server. How long is that going to take ? |
| 20:53.45 | tnt_ | huh ... can't see it @ downloads.digium.com |
| 20:54.53 | kpfleming | you didn't look very hard: http://downloads.digium.com/pub/telephony/skypeforasterisk/ |
| 20:54.55 | tnt_ | nevermind. Got redirected to downloads.asterisk.org ... |
| 20:55.16 | kpfleming | ahh, that |
| 20:55.25 | tnt_ | Yes but when I clicked on asterisk, i ended up on download.asterisk.org and from there the parent dir didn't get me back :) |
| 20:56.23 | kpfleming | yep |
| 21:02.06 | tnt_ | kpfleming: Cool, it uses RTP too, that's a perfect example :) And probably easier to follow than the 25k lines of chan_sip ... |
| 21:02.19 | kpfleming | no doubt |
| 21:02.55 | kpfleming | plus it has very little code related to actually handling the skype channels (that's all in the binary module), so it's a pretty good example of the interface between asterisk and an external network protocol |
| 21:03.53 | *** join/#asterisk-dev XoplqoX (n=chatzill@i59F50593.versanet.de) |
| 21:06.06 | tnt_ | kpfleming: which is exactly what I have too. All the 'details' are in OpenBSC itself and I just need to interface Asterisk call control to OpenBSC call control. |
| 21:06.43 | kpfleming | there you go then |
| 21:10.09 | XoplqoX | Hello, i am looking for informations howto code an channel for asterisk. I need to know which functions are called in which order. i found out that after an Dial the function ast_channel_tech.requester is called. But what happend then |
| 21:10.35 | kpfleming | we just talked about all this... find a public log of this channel and read the last 50-60 entries |
| 21:10.47 | *** join/#asterisk-dev errotan (n=errotan@62.201.123.79) |
| 21:14.03 | mvanbaak | XoplqoX: http://pastebin.ca/1581848 |
| 21:14.51 | XoplqoX | thanks |
| 21:20.02 | tnt_ | XoplqoX: As a starting point, here's what I get so far: after 'requester', that would be 'call', then at somepoint you have to sent a control message with 'answer' at that point you'll get calls to 'write' and to 'read' (assuming you setup a channel fd to monitor). Then either you'll get a 'hangup' call directly because the remote hanged up or you call ast_queue_hangup_with_cause yourself (and get called back to 'hangup' as well). |
| 21:22.05 | mvanbaak | meh, my pack of smokes tells me that smoking will make it impossible to get kids |
| 21:22.30 | mvanbaak | well, since I will never want any kids this is _NOT_ something that will make me quit smoking ! |
| 21:24.52 | mvanbaak | jsmith: how was the bday ? |
| 21:25.58 | jsmith | mvanbaak: Boring... nothing like spending your birthday away from home in a hotel... |
| 21:26.23 | jsmith | mvanbaak: I celebrated by finding a 24-hour donut shop and treating myself to a few extra calories |
| 21:28.42 | XoplqoX | the requester returns a pointer to an ast_channel which fields must be initialised in this structur to continue with the call ? |
| 21:33.20 | tnt_ | XoplqoX: I only set tech,nativeformat,readformat,writeformat and it works. |
| 21:33.49 | XoplqoX | okay i will try it thanks |
| 21:37.34 | mvanbaak | jsmith: ugh, that is boring indeed |
| 21:40.51 | *** join/#asterisk-dev CrashHD (n=CrashHD@65.74.156.109) |
| 21:59.09 | *** join/#asterisk-dev CrashHD (n=CrashHD@65.74.156.109) |
| 22:13.31 | *** join/#asterisk-dev CrashHD (n=CrashHD@65.74.156.109) |
| 23:00.27 | *** join/#asterisk-dev sruffell (n=sruffell@asterisk/the-kernel-guy/sruffell) |
| 23:00.27 | *** mode/#asterisk-dev [+o sruffell] by ChanServ |
| 23:04.22 | *** join/#asterisk-dev sruffell (n=sruffell@asterisk/the-kernel-guy/sruffell) |
| 23:04.22 | *** mode/#asterisk-dev [+o sruffell] by ChanServ |
| 23:07.20 | *** part/#asterisk-dev snuff-work (n=aaa@203.176.96.250) |
| 23:29.15 | *** join/#asterisk-dev snuff-work (n=aaa@203.176.96.250) |
| 23:29.15 | *** mode/#asterisk-dev [+o snuff-work] by ChanServ |
| 23:38.17 | *** part/#asterisk-dev XoplqoX (n=chatzill@i59F50593.versanet.de) |
| 23:39.17 | *** join/#asterisk-dev XoplqoX (n=chatzill@i59F50593.versanet.de) |
| 23:54.26 | *** join/#asterisk-dev obnauticus (n=obnautic@about/windows/regular/obnauticus) |