| 00:02.31 | *** join/#edev Crofton (~balister@ip-64-134-153-27.public.wayport.net) |
| 00:03.20 | *** join/#edev Anaphaxet0n (~george@ppp005054083139.access.hol.gr) |
| 00:05.45 | *** join/#edev prpplague (~prpplague@107-206-64-184.lightspeed.rcsntx.sbcglobal.net) |
| 00:39.32 | *** join/#edev christopher (~christoph@c-174-54-115-236.hsd1.pa.comcast.net) |
| 02:48.34 | *** join/#edev orbarron (~orbarron@192.94.92.11) |
| 04:55.39 | *** join/#edev TheSeven (~quassel@rockbox/developer/TheSeven) |
| 04:56.54 | *** join/#edev fire (~fire@unaffiliated/security) |
| 05:37.02 | *** join/#edev JoeLlama (~snork@unaffiliated/joellama) |
| 06:46.24 | *** join/#edev fire (~fire@unaffiliated/security) |
| 06:48.43 | *** join/#edev roxfan (dunno@119.222-200-80.adsl-dyn.isp.belgacom.be) |
| 09:13.47 | *** join/#edev sndcrb (~sndcrb@cs181038038.pp.htv.fi) |
| 09:28.30 | *** join/#edev pvanhoof (~philip@codeminded.be) |
| 11:49.41 | *** join/#edev g1powermac (~g1powerma@unaffiliated/g1powermac) |
| 11:54.58 | *** join/#edev utopia_euphoria (utopia_eup@nat/axis/x-eegzpynaezocrkao) |
| 12:08.47 | *** join/#edev utopia_euphoria (utopia_eup@nat/axis/x-qlegygqaxjdqvyym) |
| 12:51.31 | *** join/#edev florian (~fuchs@Maemo/community/contributor/florian) |
| 13:17.20 | *** join/#edev Anaphaxeton (~george@ppp005054083139.access.hol.gr) |
| 17:04.37 | *** join/#edev Anaphaxeton (~george@ppp089210092009.access.hol.gr) |
| 18:06.33 | *** join/#edev djerome (~djerome@ip68-2-20-108.ph.ph.cox.net) |
| 18:32.14 | *** join/#edev Crofton (~balister@32.154.102.178) |
| 19:32.38 | *** join/#edev Belliash (~belliash@asiotec/constitutive/belliash) |
| 19:49.38 | *** join/#edev Belliash (~belliash@asiotec/constitutive/belliash) |
| 20:23.08 | *** join/#edev copamp (copamp@37.250.68.62.bredband.tre.se) |
| 20:24.56 | copamp | I want to port a protocol stack from Windows to an RTOS running on an embedded system. Windows uses threads and time slicing and you can dynamically create threads. The RTOS only have static tasks and they are scheduled according to priority which means if you have 10 clients and each have a thread/task then they wont automatically be switched between, Ill have to write something that does that |
| 20:24.57 | copamp | for me. |
| 20:25.47 | copamp | So Im thinking, should I go singlethreaded or should I use tasks/threads and maybe have some token passing between each thread so they can run their stuff then let the next task run etc until they get back a token and can run again. or is that jsut apoort hack? |
| 20:27.03 | *** join/#edev itdnhr (~quassel@64.31.59.70) |
| 20:27.51 | _av500_ | what kind of embedded system? |
| 20:27.56 | _av500_ | why not run Linux on it? |
| 20:28.07 | _av500_ | is your protocoll that special? |
| 20:29.55 | copamp | well it is a gateway with 2MB RAM and 4MB Flash and a 96MHz processor. it is a proprietary OS. the protocol will also be ported to linux running on another platofmr but that is not an alternative here. |
| 20:30.09 | _av500_ | ok |
| 20:31.12 | roxfan | some rtos allow dynamic task creation too |
| 20:31.38 | roxfan | though that's not necessarily the best course here |
| 20:33.00 | copamp | well this one doesnt and i have to use this OS. |
| 20:33.01 | roxfan | you could try implementing fibers |
| 20:33.56 | copamp | so one question is, is tokenpassing between threads/tasks just a poor hack and it is perhaps better to go singlethreaded then? since it would just be a poor timeslicing mechanism to do tokenpassing. |
| 20:35.21 | *** join/#edev itdnhr (~quassel@64.31.59.70) |
| 20:37.37 | copamp | roxfan: never heard about that before, looking it up |
| 20:39.22 | roxfan | or maybe coroutines are eaiser/better here |
| 20:56.20 | copamp | roxfan: what about just going singlethreaded? |
| 20:57.01 | roxfan | well, you'll need some way of switching between different requests, no? |
| 20:57.20 | roxfan | fiber/couroutines is one way of implementing that in one thread |
| 20:57.28 | roxfan | a state maching might work too |
| 20:57.32 | roxfan | *machine |
| 20:59.37 | copamp | yes |
| 20:59.47 | copamp | ty for the advice, ill look into it |
| 21:17.25 | *** join/#edev Crofton (~balister@12.130.106.101) |
| 22:02.34 | *** join/#edev gustavoz_ (~gustavoz@host67.190-138-38.telecom.net.ar) |
| 22:03.48 | *** join/#edev gustavoz (~gustavoz@loknar.toptech.com.ar) |
| 22:06.10 | *** join/#edev Anaphaxeton (~george@ppp089210092009.access.hol.gr) |
| 22:12.16 | *** join/#edev SpeedEvil (~quassel@tor/regular/SpeedEvil) |
| 22:15.55 | *** join/#edev pvanhoof (~philip@codeminded.be) |
| 22:26.49 | *** join/#edev djerome (~djerome@ip68-2-20-108.ph.ph.cox.net) |
| 23:09.42 | *** join/#edev Anaphaxeton (~george@ppp089210092009.access.hol.gr) |
| 23:31.53 | *** join/#edev Anaphaxeton (~george@ppp089210092009.access.hol.gr) |
| 23:45.43 | copamp | roxfan: so basically what I described was cooperative multitasking? but im unsure how to implement fibers inside a task for my OS. should I just use the tasks from my OS and use mailboxes to implement cooperative multitasking? |