00:03.55 | lkcl | oh maan. udev synthesising hotplug events has _always_ killed performance. roll on depinit... |
00:10.50 | BabelOued | lkcl : any success attaching uda1380 ? |
00:10.54 | lkcl | yep. |
00:11.00 | lkcl | on the i2c bus anyway |
00:11.14 | lkcl | when i modprobed pxa-i2c suddenly things woke up oops |
00:11.20 | BabelOued | that s a good start :) |
00:11.34 | BabelOued | overrun ? |
00:12.57 | lkcl | dunno yet. usual audio rubbish stuff... GPIO... |
00:13.24 | lkcl | asoc: DAI[0:0] failed to match format |
00:13.24 | lkcl | asoc: DAI[1:0] failed to match rate |
00:13.25 | lkcl | asoc: no matching DAI found between codec and CPU |
00:13.25 | lkcl | ALSA sound/core/oss/pcm_oss.c:940: HW_PARAMS failed: -22 |
00:13.25 | lkcl | oops |
00:13.35 | lkcl | well... that's different |
00:15.11 | BabelOued | sometimes on my BA i can't run i2cdetect, it return XX for all adress :( |
00:16.29 | lkcl | ahh, that'll be due to the codec being powered off. |
00:34.33 | lamikr | Hi, I put the screenshot from newest gomunicator: http://aragorn.kortex.jyu.fi:8080/h6300/screenshots/gomunicator_front.png |
00:36.23 | lamikr | operator info is now on the right... And there is also millivolts, battery and radio link levels working. We need get the mvolt info disappear from the screen for devices not supporting it. |
00:54.33 | lkcl | Ralith: blueangel (aka xda IIs) |
00:54.55 | Ralith | ah. |
00:55.02 | lkcl | sorry. was investigating asoc. pita. mismatch between ssp driver params and codec params. |
00:55.28 | lkcl | the new asoc code (alsa system-on-chip) code is quite sophisticated... but not finished :) |
00:55.50 | lkcl | it used to be all hacked-in, into one really baaad set of code |
00:56.40 | lkcl | ok. at the _moment_ i'm on blueangel. last week it was sable. last year it was universal. year before that it was blueangel and himalaya. year before _that_ it was himalaya :) |
00:57.00 | lkcl | lamikr: cool! |
00:57.38 | lkcl | Ralith: what device you have (i missed the earlier conversation as my router got unplugged somehow) |
00:59.36 | Ralith | lkcl, I'm intending to purchase a universal soon, and very eager to have linux on it. |
00:59.40 | lkcl | cool. |
01:00.25 | Ralith | also interested in helping out if possible, although I'm not really very experienced in C, let alone reverse engineering/kernel hacking |
01:01.24 | lkcl | well, the universal is the closest thing to 'complete' of the HTC phones. |
01:02.16 | Ralith | which is one of the reasons I'm going for it instead of, say, the TyTN |
01:03.00 | lkcl | if you wanted to help out, you could always set up an openembedded.org build environment, and build your own opie+familar and gpe distributions, and then publish them. |
01:03.04 | lkcl | heh. |
01:03.15 | lkcl | well, the TyTN (codenamed hermes) is next on the list. |
01:04.13 | Ralith | isn't there already a published opie+familiar distro? |
01:04.58 | Ralith | (also, the Universal is much cheaper on ebay than the TyTN, and the full VGA screen + bigger keyboard are both things I look forward to eagerly) |
01:05.27 | lkcl | yerss.... published, yes, not-necessarily-up-to-date-and-specifically-tailored-to-the-universal, no |
01:05.44 | lkcl | but speak to people here as they drop by, and ask if that would be something useful. |
01:06.13 | lkcl | ahh, but the universal's battery life (due to the pxa27x absolutely shit power management) kills it for me. |
01:06.39 | lkcl | i much prefer the hermes as it has the samsung s3c2442 processor. 6 day battery life |
01:06.42 | Ralith | can't a software-level cpufreq scaler help? |
01:07.26 | Ralith | also, I don't intend to be away from my charger for more than 12 hours |
01:09.08 | Ralith | thoughts? |
01:10.33 | lkcl | it's the suspended (0Hz) state that has the awful power management on the intel arm processors. |
01:11.02 | Ralith | how long does it take to boot? |
01:11.03 | lkcl | at least the Texas Instruments OMAP processors got it right, although those are _really_ bad processors capable of at most 200mhz |
01:11.24 | Ralith | into, say, the opie gui |
01:11.33 | lkcl | the universal? probably about... 1 to 2 minutes, at a guess. |
01:11.46 | Ralith | hm. |
01:11.51 | lkcl | it's a 500mhz processor, equivalent to a 1ghz Pentium III |
01:11.58 | Ralith | so there's a good reason to suspend |
01:12.38 | lkcl | but, in user-space, there is udev, which takes a good 20 seconds of that boot time just sitting there (i know why - it's complicated to explain: look up depinit advogato on google) |
01:12.58 | lkcl | http://www.advogato.org/person/lkcl/diary.html?start=280 |
01:13.14 | Ralith | is that something that can be overcame? |
01:13.18 | lkcl | 10 may entry |
01:13.19 | lkcl | yes. |
01:14.06 | lkcl | but like i said, it's complicated to explain. you need a parallel init (startup / process management) system to get round the problem. |
01:14.13 | Ralith | is there any chance of that being done soon? :P |
01:14.33 | Ralith | is that likely to be implemented? |
01:14.46 | lkcl | so - if you _really_ wanted a task-and-a-half, then you could try compiling up depinit for your shiny new universal you're about to get :) |
01:14.58 | Ralith | :P |
01:14.59 | lkcl | i can send you my scripts that i have running on my fujitsu laptop. |
01:15.06 | lkcl | are you any good at shell scripts etc? |
01:15.18 | Ralith | I know the basics of bash and can learn. |
01:15.57 | lkcl | the boot-time would be cut to about under 30 seconds, using depinit, and power-down time to under FIVE seconds. |
01:15.57 | Ralith | cool. |
01:15.58 | lkcl | sysvinit is shit. |
01:16.24 | lkcl | http://www.nezumi.plus.com/depinit/index.html |
01:16.47 | Ralith | and yer sure that depinit is the best option? |
01:16.56 | lkcl | i think that the arm processors of 400mhz + can be considered 'modern' by now |
01:17.51 | lkcl | ok - there _are_ alternatives: there is gentoo's parallel-startup system, which is based on the principles of depinit, because depinit was in linux-from-scratch, and gentoo started from linux-from-scratch with portage (ported from freebsd to linux) |
01:17.53 | lkcl | whew :) |
01:18.21 | lkcl | but i _understand_ depinit, and i know the author. so i can help advise you. |
01:18.27 | Ralith | heh |
01:18.31 | Ralith | works for me |
01:18.37 | lkcl | good man! |
01:19.07 | lkcl | if you get it started, that will at least encourage me to kick in and help |
01:19.21 | lkcl | anyway. i got to sort out sound. |
01:19.32 | Ralith | good luck |
01:20.00 | Ralith | oh, one more random question, if you have time: what's wince's boot time, for comprison? |
01:21.26 | BabelOued | lkcl: if you want to find where time is lose : http://www.bootchart.org/ |
01:26.12 | *** join/#htc-linux BHSPitMonkey (n=stephen@adsl-65-68-200-65.dsl.rcsntx.swbell.net) |
01:26.14 | lamikr | lkcl: How is the power management handled in htc devices. Do you use apm kernel emulation or "echo mem /sys/power/state" way of doing things... |
01:26.20 | BHSPitMonkey | good afternoon |
01:29.49 | BHSPitMonkey | I hate to be so inappropriately appropriate, but I figured here would be my best shot: is there a way to share internet access with my smartphone via bluetooth? (phone=wm5.0, pc=linux) |
01:30.06 | Ralith | yeah |
01:30.16 | Ralith | well |
01:30.18 | Ralith | which way? |
01:31.25 | lkcl | taaa |
01:31.57 | Ralith | the phone can be made to act as a modem |
01:32.08 | Ralith | it's been done before |
01:32.14 | Ralith | google will find a fair number of results |
01:32.17 | lkcl | lamikr: i just do apm --suspend, from opie |
01:32.18 | BHSPitMonkey | yeah, I meant the reverse |
01:32.51 | Ralith | BHSPitMonkey, it should be possible, but I dunno how. |
01:33.16 | lkcl | BabelOued: my fujitsu laptop, running debian with depinit, takes the usual 20 seconds to start all the kernel, and the first process, instead of /sbin/init, is /sbin/depinit. from there, it is under 20 seconds to start Xorg and _all_ other services. |
01:33.17 | BHSPitMonkey | I used to do it with my symbian phone |
01:33.45 | lkcl | the same laptop: if i use /sbin/init, then /sbin/init takes OVER NINETY seconds to get everything started. |
01:34.45 | lamikr | lkcl: Ok, with omap kernels the apm emulation is not supported.--> will hang. Instead we use this "/sys/power/state" way. I am just missing an userspace app like apmd for this other way of doing things which could but my device to sleep after couple of minutes inactivity. |
01:34.48 | lkcl | sysvinit was designed for when context-switching took forever, because the maximum CPU speed available, on a 68020 or 68030, was 16mhz if you were lucky. |
01:34.58 | BabelOued | lkcl: on suse we have same, it is startpar :) and bootchart help me to optimise a little between io usage and cpu usage |
01:35.24 | lkcl | startpar is a parallel ok cool. very good. |
01:36.05 | lkcl | depinit is sufficiently advanced that it actually finds bugs in the linux kernel :) |
01:36.36 | Ralith | haha |
01:36.36 | Ralith | nice |
01:36.55 | Ralith | sounds impressive |
01:37.07 | lkcl | you couldn't use it for early 2.6 kernels because signals would not be correctly sent to process 1, if process 1 had been put into 'stop' state rather than 'running' :) |
01:38.06 | lkcl | the solution was to have a script which constantly sent a kill -HUP to process 1, every second. that kept process 1 'awake' and so it received all the other signals from all (parent - i.e. eeevery) processes. |
01:38.12 | lkcl | it's a long story. |
01:38.40 | lkcl | but Ralith: you would do well to investigate e.g. startpar, and the gentoo one, and depinit etc. etc. |
01:39.22 | Ralith | -_- |
01:39.25 | lkcl | which brings me on to another - different - point: kernel-driver parallel-initialisation |
01:39.30 | Ralith | sounds like an obscene amount of work |
01:39.38 | Ralith | and how likely is that to get implemented? |
01:39.47 | lkcl | well, it took me two weeks to get depinit installed on my debian laptop. |
01:40.23 | lkcl | basically, i was working from richard lightman's customised (and very sophisticated) linux-from-scratch depinit example scripts, to create debian-ish ones. |
01:40.39 | lkcl | long story. |
01:40.41 | Ralith | best to use custom ones anyway, in a context like this |
01:41.22 | lkcl | well, they're pretty much customised anyway: see /etc/init.d on the opie distro for example. |
01:41.54 | Ralith | so what's this about parallel driver init? |
01:42.13 | lkcl | but basically, for every entry in /etc/init.d/ you need to create a matched corresponding pair of /etc/depinit/<name>/startup and stop scripts or... well, that's not quite true: it's a long story |
01:42.15 | lkcl | ah. |
01:42.20 | lkcl | well... |
01:42.51 | lkcl | if you have two busses, what the hell are you waiting around to initialise one bus - and its devices - followed subsequently by then starting on the second bus?? |
01:43.23 | Ralith | explain depinit to me once I have the universal (which might not be too soon) |
01:43.24 | Ralith | :P |
01:43.32 | Ralith | but yeah |
01:43.35 | lkcl | that's the basic question. |
01:43.36 | Ralith | I get why it's a good idea |
01:43.45 | Ralith | but how likely is it to be done? |
01:43.59 | lkcl | platform_device(s) are now set up as 'children'. |
01:44.07 | lkcl | the asic3_base.c is a good example. |
01:44.17 | lkcl | so in theeeoory, it should be possible to do :) |
01:44.26 | Ralith | yeah, but who's going to? :P |
01:44.38 | lkcl | well... me, probably. as an experiment. |
01:45.00 | Ralith | heh |
01:45.01 | lkcl | but, to be honest, udev with modprobe does most things in parallel, anyway. |
01:45.02 | Ralith | yay! |
01:45.08 | Ralith | even so. |
01:45.11 | Ralith | good to be complete |
01:45.20 | lkcl | it's just that if you compile your kernel with everything built-in... |
01:45.29 | Ralith | and how far-ranging will be these changes? |
01:45.31 | lkcl | so, udev takes care of _most_ of it. |
01:45.35 | lkcl | i honestly don't know. |
01:45.36 | Ralith | that phone, HTC phones, or all of linux? |
01:45.41 | Ralith | not in terms of code :P |
01:48.39 | lkcl | all of linux. |
01:48.44 | lkcl | it's only a thought, so far. |
01:48.48 | Ralith | cool. |
01:48.50 | Ralith | good luck :) |
01:49.05 | lkcl | it _might_ be something as simple as putting in a call to schedule_work() somewhere in platform_device_register() |
01:49.31 | lkcl | anyway. sleep time soon. |
01:53.22 | Ralith | seeya. |
02:02.34 | *** join/#htc-linux ellisway (n=ellis@host-87-74-241-174.bulldogdsl.com) |
02:10.38 | *** join/#htc-linux ralith_ (n=ralith@soggy202.drizzle.com) |
02:27.58 | *** join/#htc-linux lkcl_ (n=lkcl@535411CB.cable.casema.nl) |
04:18.38 | *** join/#htc-linux psokolovsky_ (n=psokolov@ip.85.202.124.214.dyn.sub-9.broadband.voliacable.com) |
06:11.20 | *** join/#htc-linux JTRipper86 (n=jtripper@p54B1D52A.dip.t-dialin.net) |
07:06.47 | *** join/#htc-linux goxboxlive (n=goxboxli@gw.hjelcowi.no) |
07:14.39 | *** join/#htc-linux LunohoD (n=alex@e180064146.adsl.alicedsl.de) |
07:25.55 | *** join/#htc-linux gw280 (i=authdeni@vm.gwright.org.uk) |
07:59.52 | *** join/#htc-linux rmoravcik (n=rmoravci@pc-3s0zt5w2e4y0vzmhnrzq3a21zqajzfw.users.student.utc.sk) |
08:09.33 | *** join/#htc-linux Fossi (n=cmewes@195.177.48.115) |
09:30.32 | *** join/#htc-linux bd2 (n=ufo@89.113.84.151) |
10:07.23 | *** part/#htc-linux goxboxlive (n=goxboxli@gw.hjelcowi.no) |
12:08.44 | *** join/#htc-linux WizMaui (n=WizMaui@62.112.90.231) |
12:29.11 | *** join/#htc-linux rob_w (n=bob@p85.212.155.144.tisdip.tiscali.de) |
13:16.05 | *** join/#htc-linux TripleDES (n=sergio@89.129.143.241) |
13:22.19 | *** join/#htc-linux rmoravcik (n=rmoravci@pc-3s0zt5w2e4y0vzmhnrzq3a21zqajzfw.users.student.utc.sk) |
13:24.49 | *** join/#htc-linux psokolovsky_ (n=psokolov@ip.85.202.124.214.dyn.sub-9.broadband.voliacable.com) |
13:51.16 | *** join/#htc-linux skodde (n=skodde@unaffiliated/skodde) |
14:33.15 | *** join/#htc-linux skodde (n=skodde@unaffiliated/skodde) |
15:17.49 | *** join/#htc-linux skodde (n=skodde@unaffiliated/skodde) |
15:24.04 | psokolovsky_ | Hi! |
15:24.49 | psokolovsky_ | I sent a report regarding OPIE's issues in Angstrom and way to solve the situation to htc-pxa@ list. |
15:25.01 | psokolovsky_ | expecting your response, guys ;-) |
15:56.11 | *** join/#htc-linux rob_w (n=rob_w@p549B93C5.dip0.t-ipconnect.de) |
15:58.51 | *** join/#htc-linux Ralith (n=ralith@66.150.12.86) |
16:45.48 | *** join/#htc-linux ellisway (n=ellis@host-87-74-241-174.bulldogdsl.com) |
16:58.57 | *** join/#htc-linux gw280 (i=authdeni@vm.gwright.org.uk) |
17:02.49 | *** join/#htc-linux goxboxlive (n=goxboxli@9.80-202-160.nextgentel.com) |
17:11.49 | *** join/#htc-linux Ralith (n=ralith@66.150.12.86) |
17:14.13 | *** join/#htc-linux ralith_ (n=ralith@66.150.12.86) |
17:22.40 | goxboxlive | hi |
17:23.11 | goxboxlive | psokolovsky_ So you think that we shal make a new distr. |
17:23.39 | goxboxlive | There are AFAIK two issues with angstrom and opie. |
17:23.49 | goxboxlive | 1) Is the missing getkey |
17:24.06 | psokolovsky_ | goxboxlive, no, that's just an idea ;-) but I lean to it, after all, assuming we'll be able to work together on it. |
17:24.15 | Ralith | you can just fork the existing ones, can't you? |
17:24.18 | Ralith | add what you need |
17:24.20 | goxboxlive | 2) Is that psplash screwing up the opie gui |
17:25.00 | goxboxlive | I think i have solved the psplash stuff. |
17:25.45 | goxboxlive | In /etc/init.d/opie there is a line that days this: echo Starting Opie.... |
17:25.45 | goxboxlive | <PROTECTED> |
17:25.54 | psokolovsky_ | goxboxlive, believe me, those are only first and relatively minor issues ;-) |
17:26.16 | psokolovsky_ | goxboxlive, but you're right that with own distro we would be free to solve them as we like ;-) |
17:26.33 | goxboxlive | If we move it to terminal 3 i guess it will work. Seems like bootup takes tty1, psplash tt2 and also opie tty2. |
17:27.20 | psokolovsky_ | goxboxlive, but the real big issue OPIE, GPE, and anything else has in OE (and any other such system) is inconsistent device support. That's what I'd be interested to address. And there will be lots of breakage on that way ;-) |
17:27.25 | goxboxlive | Yes, we could just as you said make a new one based on Angstrom |
17:27.32 | psokolovsky_ | goxboxlive, Sounds scary, right? ;-) |
17:28.12 | goxboxlive | yes :-) |
17:28.12 | psokolovsky_ | goxboxlive, good analysis on psplash issue |
17:28.36 | psokolovsky_ | goxboxlive, so, please give it good thought, discuss with cr2/other folks, and then let's see what we'll have... ;-) |
17:29.23 | goxboxlive | Yes, of course. |
17:31.46 | goxboxlive | But if you do that, i think we just shall just consentrate on Opie since there are allredy a distro that has fully GPE support. |
17:32.05 | goxboxlive | s/you/we/ |
17:34.38 | psokolovsky_ | goxboxlive, sure, only opie. angstrom will do gpe ;-) |
17:34.39 | goxboxlive | 22222222Ralith: Yes you can fork the exiting ones, but the issue here is that Angstrom will not support Opie, only GPE. So at the moment the Angstrom-opie-image doesnt work quite right. You have to do some hacks to get it working right. |
17:34.46 | goxboxlive | psokolovsky_ right. |
17:37.25 | cr2 | goxboxlive: can you explain me why opie support is such a big issue (in technical sense) ? |
17:37.58 | goxboxlive | cr2: Do you mean from our suide of view or that Angstrom wouldnt support it? |
17:38.08 | cr2 | i have created http://wiki.xda-developers.com/index.php?pagename=HTC_Audio |
17:38.16 | cr2 | goxboxlive: from our side |
17:38.46 | cr2 | what is so special about angstrom with gpe removed ? |
17:39.12 | goxboxlive | cr2: Well, the Angstrom distro is mor up 2 date than familiar. |
17:39.31 | cr2 | think angstrom+openmoko, or angstrom+x11+qt |
17:39.54 | goxboxlive | but i think angstrom+x11+qt will be to slow |
17:40.09 | cr2 | why ? |
17:40.24 | cr2 | let's be more specific. |
17:40.33 | cr2 | can i remove gpe from angstrom ? |
17:41.00 | cr2 | remove == exclude all gpe-dependent packages. |
17:41.01 | goxboxlive | Yes, you are able to make a angstrom-bootstrap-image |
17:41.12 | cr2 | ok. |
17:41.16 | goxboxlive | And that should be gpe free. |
17:41.26 | cr2 | can we build a eabi opie image. |
17:41.51 | cr2 | afaik it's located completely in /opt/Qt* |
17:42.01 | goxboxlive | isent it eabi in angstrom? |
17:42.13 | cr2 | does it need some additional packages ? |
17:42.21 | goxboxlive | i guess my compillator is eabi |
17:42.38 | cr2 | eabi is binary convention in a sense. like elf/ecoff. |
17:43.23 | cr2 | ok, so we can build eabi opie |
17:43.42 | cr2 | are there eabi-related opie problems ? |
17:44.35 | cr2 | keyboard is not eabi-related :) |
17:45.21 | goxboxlive | no, there arent, But it is the getkey support, and if there will be any changes in angstrom later they wouldnt take care about opie issues |
17:45.39 | cr2 | which package is getkey ? |
17:45.48 | goxboxlive | NOt that we SO need getkey. But that's the first issue. |
17:46.02 | goxboxlive | There is no getkey package, it is a patch to busybox |
17:46.21 | cr2 | busybox,i guess. does busybox support stty 921600 now ? |
17:46.47 | cr2 | if not, we need to patch it too. |
17:46.55 | goxboxlive | i dont know, i havent tried |
17:47.05 | cr2 | so getkey+921600 pathes are needed. |
17:47.16 | cr2 | s/pathes/patches/ |
17:47.22 | cr2 | we need it on sable. |
17:47.43 | goxboxlive | I dont know about 921600. Ask psokolovsky_ if they have have support for it. |
17:47.54 | cr2 | we will resolve the keyboard issue. |
17:48.05 | cr2 | sable is the only device is know that needs it. |
17:48.21 | goxboxlive | ok |
17:49.00 | cr2 | the zroadmap patch from your collection you can submit to oe anyway. |
17:49.44 | cr2 | maybe they don't care, but it's an openzaurus package, so it should be accepted. |
17:51.25 | goxboxlive | yes, i know. But that's not a porblem i think. But the problem is as i said when changes in Angstrom distro they wouldnt care about breaking some opie stuff. And that will end up with a useless opie image. Nad when no one is able to build opie then it will be remved form OE.DEV |
17:52.13 | goxboxlive | Because, if you use familar now, opie wouldnt build. You have to do some hack to build it. |
17:53.56 | goxboxlive | And if we made a own distro based on the angstrom, but optimized for Opie we will have both gpe (with agnstrom) working, and Opie working with a Angstrom based distro. |
17:54.25 | *** join/#htc-linux gw280 (i=authdeni@vm.gwright.org.uk) |
17:54.57 | goxboxlive | But, maybe we shouldnt do this before we have tried out Openmoko. If Openkoko is a killer then i guess no one will use Opie/GPE anyway. |
17:55.50 | goxboxlive | Beacuse i think Angstrom will support Openmoko since Mickey isone of the Openmoko gyus and he also is a Angstrom developer. |
17:56.49 | cr2 | i want zroadmap, and it's a qt app :) so i do care. |
17:57.04 | cr2 | i think this is all too abstract talk. |
17:57.37 | cr2 | if this "distro" will differ by 1k patches, i don't really care. |
17:57.59 | cr2 | and don't tell me getkey breaks opie :) |
17:58.14 | cr2 | i just don't see how angstrom can break opie. |
17:58.35 | goxboxlive | No, it doesne break opie. But as i said, what's next |
17:58.55 | cr2 | well, what _can_ break opie ? |
18:00.30 | cr2 | i will not put too much hope in openmoko in near future, because it does not have bt,wifi,umts,videophone. |
18:00.30 | goxboxlive | Well i guess it want break Opie, but what about if they include some GTK+ shit, and building angstrom will break without it. |
18:00.40 | goxboxlive | i c |
18:01.06 | cr2 | you have told me there is angstrom-bootstrap. |
18:01.30 | cr2 | let's think what we need on top of angstrom-bootstrap. |
18:02.28 | goxboxlive | hmm, yes there is a angstrom-bootstrap-image |
18:02.32 | cr2 | if they want to break opie intentionally, they can always do it. but i doubt it's a very big issue, because this breakage will not be big. |
18:03.19 | cr2 | and they are not microsoft, their code is open, it is easy to fix the breakage. |
18:05.26 | cr2 | i don't really think the running something off the framebuffer is a very bright idea nowdays, that's why i'd always like to have a x11+qt option. |
18:05.51 | goxboxlive | i c |
18:08.03 | goxboxlive | Well i dont know what to do, but i understand that psokolovsky mention it, because stuff that he would like to add or fix in favour of Opie wouldnt be accepted, and that's kind of frustrating for him i guess since he is the only one who maintain Opie in OE. |
18:10.12 | psokolovsky_ | goxboxlive, well, I'm not the only one who works on opie, but I'm the one who took duties it will be kept in OE, and will serve purpose for users. So, current movement is more based on the issues you experience right now, then myself. |
18:10.41 | psokolovsky_ | (As you know, I work on many things, and don't have enough time right now to pay to opie) |
18:10.58 | psokolovsky_ | That's why I ask you to think well what route youw nat to take - |
18:10.59 | goxboxlive | ok |
18:11.40 | psokolovsky_ | I don't want to lead you into something you didn't really want, nor want to work on something which won't help opie as a whole ;-) |
18:12.57 | *** join/#htc-linux BabelOued (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net) |
18:14.26 | cr2 | psokolovsky_: what _real_ problems do we have with opie now ? |
18:15.23 | psokolovsky_ | cr2, "There's no distro which supports contemporary OPIE." |
18:15.24 | goxboxlive | You right. Well i dont know what's best to do. For now, there are only thoose two issues i know about. And both of them are solved. I think we shall remove getkey form opie, and cgange the opie start to tty3. |
18:16.04 | psokolovsky_ | goxboxlive, I'm glad about your decision regarding getkey |
18:16.07 | goxboxlive | And i have just asked koen about angstrom support of Openmoko, and they will support it as long as it's not dead as Opie. |
18:16.34 | psokolovsky_ | ok |
18:16.49 | cr2 | what distro support do we need/expect ? |
18:17.01 | BabelOued | hi everybody |
18:18.14 | cr2 | making sure that opie works with standard linux keyboard is an opie internal problem, not distro-specific. |
18:18.53 | psokolovsky_ | cr2, In my view, there should be distro which would support OPIE 1.2.2 in such way taht it is usable after install, and for as much devices as possible (and the key to that would be supporting different devices consistently and uniformly) |
18:19.42 | cr2 | ok, but isn't it more a problem of the kernel interface ? |
18:19.56 | psokolovsky_ | cr2, at this time, you may have hard time knowing what OPIE-specific problems exist at all, as you can't get arbitrary device and install QAed release of OPIE 1.2.2 on it |
18:20.33 | psokolovsky_ | cr2, yeah, good catch ;-) exactly, what should be done on kernel level, should be done there, not on userspace level ;-) |
18:20.33 | cr2 | ok. how many devices do we have in active development ? |
18:21.34 | psokolovsky_ | that's essense of my sly plan - drop all the gory crap out of qte, libopie2 and force kernel maintainers of individual devices to their homework. Bwa-ha-ha!!! ;-) |
18:21.51 | psokolovsky_ | cr2, we can count 10 easily |
18:22.25 | cr2 | it seems to me that full hx4700+universal support will cover all possible PDA/phone options |
18:22.43 | cr2 | i have a feeling that it's more close to 5. |
18:23.05 | cr2 | really different systems, that need attention. |
18:23.13 | psokolovsky_ | cr2, if you want to call it that way - "hx4700+universal". I'd call it "normal PDA support + extended phone features" |
18:23.38 | cr2 | i count all htc pxa phones as 1 device. |
18:23.41 | cr2 | agreed. |
18:23.46 | psokolovsky_ | cr2, well, *all* PDAs are just the same. so, 1 system ;-P. phone stuff is extra, not difference. |
18:24.12 | cr2 | hx4700 and universal as a pda are mostly exact clones. |
18:24.58 | cr2 | it makes some difference if you have pxa or s3c24xx. |
18:25.02 | cr2 | or omap. |
18:25.03 | psokolovsky_ | Any PDA has LCD, backlight, CPU with need for suspend/freq ctrl, buttons, memory card slots. |
18:25.13 | psokolovsky_ | Some PDAs have BT, WiFi, kbd, etc. |
18:25.24 | psokolovsky_ | - any PDA will fit that description |
18:25.57 | psokolovsky_ | and differences between 2 different kbd's would rather be handled on kernel level as mush as possible. |
18:26.07 | cr2 | yes, but the cpu is a main "building block". |
18:26.14 | cr2 | yes. |
18:26.23 | psokolovsky_ | cr2, CPU diff is also on kernel level, right? ;-) |
18:26.30 | cr2 | yes. |
18:26.33 | cr2 | :) |
18:26.56 | cr2 | but if you have /sys/class/pxafb it makes a difference. |
18:27.40 | psokolovsky_ | cr2, don't rely on pxafb, rely on /sys/class/lcd/*/... ;-) |
18:27.47 | cr2 | yes. |
18:27.52 | psokolovsky_ | unless we talk about hw accel, but that's another matter |
18:28.12 | cr2 | it's great which works you have done unifying all asic3-dependent devices. |
18:28.15 | psokolovsky_ | it's also like "some PDAs have hw accel for X, but all that handled consistently" |
18:28.35 | psokolovsky_ | cr2, we all do that, and there's even more to do ;-) |
18:28.37 | cr2 | very little current pdas have hw accel |
18:28.45 | cr2 | true. |
18:29.03 | psokolovsky_ | but yes, the picture as described is what I have in mind as the aim ;-). a bit distant, but worthy, imho |
18:29.42 | cr2 | agreed. |
18:31.09 | *** join/#htc-linux rob_w (n=bob@p85.212.187.248.tisdip.tiscali.de) |
18:32.08 | *** join/#htc-linux cr2 (n=konversa@crpl22.physik.uni-wuppertal.de) |
19:30.06 | BabelOued | lkcl_: are you here ? |
19:30.30 | BabelOued | ba doesn't boot with latest kernel :( black screen |
19:35.04 | *** join/#htc-linux Fossi (n=Fossi@e176126097.adsl.alicedsl.de) |
19:39.30 | *** join/#htc-linux asylumed (n=insanity@196.211.28.91) |
19:44.26 | BabelOued | lkcl_: it is problem with latest w100fb patch :( |
20:03.13 | goxboxlive | cr2: When i am using gpe and hit the suspend icon it suspends. It also resume in to GPE again. |
20:10.10 | *** join/#htc-linux gw280 (i=authdeni@vm.gwright.org.uk) |
20:29.04 | cr2 | goxboxlive: ok. it's obviuosly the second annoying opie bug. |
20:32.21 | goxboxlive | :-) |
20:46.36 | *** join/#htc-linux ellisway (n=ellis@host-87-74-241-174.bulldogdsl.com) |
20:59.34 | *** join/#htc-linux WizMaui (n=WizMaui@62.112.90.231) |
21:07.04 | *** join/#htc-linux RoEn_PC (n=roen@p54A66A2E.dip.t-dialin.net) |
21:11.14 | psokolovsky_ | BabelOued, did you try that it's exactly latets patch to w100fb? If so, please report that to kernel-dicuss@. |
21:17.00 | BabelOued | psokolovsky_: it is revision 1.29 "Add Manuel Tiera's patch" |
21:17.13 | psokolovsky_ | BabelOued, please report. |
21:17.25 | BabelOued | maybe it need some new initialisation, i m looking at |
21:17.46 | psokolovsky_ | ok, fixing it would be indeed better ;-) |
21:18.01 | psokolovsky_ | I asume Matt tested it on hx4700 before committing... |
21:18.52 | BabelOued | psokolovsky_: ok so i look at hx4700 to see if there is new initialisation to add into blueangel |
21:32.05 | *** join/#htc-linux WizMaui_ (n=WizMaui@62.112.90.231) |
21:54.27 | *** join/#htc-linux gw280 (i=authdeni@vm.gwright.org.uk) |
21:59.37 | goxboxlive | ohh no, my universal just died.wtf |
22:02.55 | goxboxlive | hmm it waked up, strange. I have allmost 80% battery. The powerled turned red and it was all dead, beside when i removed the usb cable the red led shut off. But when i plugged the cable again it turned red. |
22:03.30 | goxboxlive | well, it's working now, powerled are again orange |
22:15.30 | *** join/#htc-linux BabelOued_ (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net) |
22:54.59 | *** join/#htc-linux gw280 (i=authdeni@vm.gwright.org.uk) |
23:11.17 | *** join/#htc-linux florian (n=fuchs@87.193.38.108) |
23:42.04 | *** join/#htc-linux awelux (n=awelux@lvps87-230-8-217.dedicated.hosteurope.de) |
23:45.35 | *** join/#htc-linux awelux (n=awelux@lvps87-230-8-217.dedicated.hosteurope.de) |
23:52.18 | *** join/#htc-linux awelux (n=awelux@lvps87-230-8-217.dedicated.hosteurope.de) |