00:00.29 | cjb | Actually, discovery of the day is that python runs 40% slower when built with --enable-shared. Are symbol lookups in shared libraries known to go pathological with low cache? |
00:00.38 | lkcl | try running prelink, on the actual laptop. i have a via 500mhz mini-itx board and it takes TEN HOURS to run prelink on it - compared to 10 minutes on a 1ghz processor |
00:00.47 | lkcl | you know the man to _really_ ask that question of? |
00:00.57 | cjb | Nope, but I'd like to. :) |
00:01.02 | lkcl | andrew tridgell. |
00:01.06 | lkcl | tridge@samba.org |
00:01.43 | cjb | Huh. Well, okay. I might save his time by trying the python lists first, but I'll certainly drop him a mail if we don't figure this out. |
00:01.59 | lkcl | he will be able to explain to you in seriously-sufficient detail. |
00:02.04 | cjb | Neat. |
00:02.19 | cjb | OOI, what arch was the CPU you designed? |
00:02.27 | lkcl | ok - to explain (hope everyone else doesn't mind the OT...) |
00:02.40 | lkcl | oh - a micro-code nano-kernel thing. |
00:02.44 | lkcl | it's complicated. |
00:02.46 | cjb | Gosh. |
00:02.51 | cjb | Got a paper or anything? |
00:03.01 | lkcl | an asynchronous processor design in under 4,000 gates. |
00:03.20 | lkcl | so little silicon that you can put 32 or 64 of them on a single die |
00:03.34 | lkcl | naah. |
00:03.44 | lkcl | lots of emails and skype conversations, though. |
00:04.38 | lkcl | the original designer did a processor, in 1992, that ran at over 200mhz, in CMOS, when everyone else was struggling to reach 90mhz. |
00:05.03 | lkcl | it was an intel 8046 clone - specialist medical processor. |
00:05.26 | cjb | Sounds fun. |
00:05.28 | lkcl | and the designer still has his notes, and the rights to do a redesign. |
00:05.33 | lkcl | oh - absolutely fantastic. |
00:06.27 | lkcl | but absolutely _not_ going to happen, without any funding. the people we approached even have their own fab and cadence etc. licenses (they do flash memory chips and processor clones) |
00:07.17 | cjb | Suck. Gotta hate it when business gets in the way of fun technology. :) |
00:07.32 | lkcl | in 90nm, we could easily push 1ghz, and possibly even 2ghz. 2ghz micro-code processors, times 32. and if they're not busy, they use zero power, automatically, because they're all asynchronous. |
00:08.19 | lkcl | yeh they split the company - a 200-million-dollar company - over the row about $20m. |
00:08.25 | lkcl | anyway. |
00:08.29 | lkcl | _where_ were we? :) |
00:08.39 | cjb | heh. something about the Sable vs. the Universal. |
00:08.44 | lkcl | oh yes. |
00:08.46 | lkcl | right :) |
00:09.01 | lkcl | it's up to you, basically. compare these two pages: |
00:09.03 | lkcl | 1sec... |
00:09.47 | lkcl | http://wiki.xda-developers.com/index.php?pagename=Ipaq6915 |
00:10.35 | lkcl | http://wiki.xda-developers.com/index.php?pagename=UniversalResearch and also the Technical page |
00:10.50 | lkcl | ok maybe not the technical page |
00:11.05 | cr2 | http://www.handhelds.org/moin/moin.cgi/Universal |
00:11.24 | lkcl | and then this: |
00:11.26 | lkcl | http://wiki.xda-developers.com/index.php?pagename=UniversalProgress |
00:11.26 | cr2 | i think this one is more up to date. for non-tech stuff |
00:11.42 | BabelOued | cr2: what is command for cu ? |
00:11.52 | lkcl | against this: |
00:11.54 | lkcl | http://wiki.xda-developers.com/index.php?pagename=SableProgress |
00:12.04 | lkcl | sorry the sableprogress page isn't as nice as the universal one. |
00:12.23 | cr2 | BabelOued: cu -l /dev/ttyS? -s 115200 |
00:12.44 | cr2 | http://wiki.xda-developers.com/index.php?pagename=HTC_Serial |
00:12.53 | cr2 | on blueangel /dev/ttyS0 |
00:13.21 | BabelOued | thanks |
00:13.44 | lkcl | sable: simplified hardware, in my opinion stands a better (easier?) chance of completion than the universal. but the universal has cr2, who is a persistent sod who just won't give up because he likes the universal. |
00:13.54 | cjb | lkcl: is it easier to get hold of either of the phones? |
00:14.06 | lkcl | cjb: yep. ebay is your friend. |
00:14.31 | lkcl | 16k cache. shit. |
00:14.34 | cr2 | http://www.handhelds.org/moin/moin.cgi/UniversalStatus |
00:14.51 | lkcl | i can't believe you went for a processor with only 16k cache. |
00:15.05 | cr2 | hmm. a bit outdated too. |
00:15.08 | cjb | lkcl: we wanted x86 and FPU and low-power. |
00:15.23 | lkcl | you should have gone for the opencores.org 1200 processor, instead. |
00:15.41 | lkcl | the only advantage to x86 is the amazingly compact instruction set. |
00:16.02 | lkcl | which is only offset by the amazingly long time it takes the processor to decode the bloody instructions. |
00:16.30 | lkcl | hence an arm processor at 500mhz is equivalent to a 1ghz pentium. |
00:16.46 | cjb | lkcl: we want to ship 100 million laptops/year in a few years' time, so also need someone who can make that many cheaply. |
00:16.47 | lkcl | ok - only _significant_ technical advantage... |
00:17.03 | cjb | (the world makes 45 million laptops/year at the moment. in total.) |
00:17.42 | lkcl | well arm processors _far_ outnumber x86 processors by ... i wouldn't be surprised if it's probably an order of magnitude, because they are used in nearly every mobile phone on the planet! |
00:18.19 | cjb | Yeah, ARM is everywhere. Not usually with an FPU, though. |
00:18.32 | lkcl | but yes - the whole reason why smartphone / pda designers picked the pxa2xx series is because it could say 'intel inside' and people would be reassurred. |
00:19.12 | lkcl | the dec/alpha had such a good 32-bit multiplier and such an efficient instruction set that it didn't _need_ an FPU. |
00:19.45 | cjb | yeah, plenty of ex-alpha people here. |
00:19.49 | lkcl | if you have a chip with a 1s complement add instruction (i think it's add) and a sign-extend instruction, you don't need an FPU. |
00:19.56 | lkcl | well then! :) |
00:20.34 | lkcl | my friend richard lightman knows a lot about processor designs - but it would be good to have another source confirm what he told me |
00:21.05 | lkcl | anyway. |
00:21.11 | lkcl | off-topic again, ha ha |
00:21.17 | lkcl | phones. right. |
00:21.29 | cr2 | lkcl: http://lists.openezx.org/pipermail/openezx-devel/2006-December/000865.html |
00:21.38 | lkcl | ta cr2... 1sec... |
00:21.49 | BabelOued | cr2: it is hard to prefix each at command in cu ! :( |
00:21.51 | lkcl | btw gotta go sleep soon.... |
00:22.00 | cr2 | that may be the reason why ba/hima have a ssp/i2s split |
00:22.21 | cr2 | BabelOued: use 'chat' |
00:23.13 | lkcl | cr2: o yuk |
00:23.48 | lamikr | cr2: the bts parser works nicely. |
00:24.00 | lkcl | cr2: yep. |
00:24.30 | cr2 | lamikr: i have modified it to extract the ogf/ocf |
00:24.59 | cr2 | lamikr: http://wiki.xda-developers.com/index.php?pagename=UniversalBluetooth |
00:25.01 | lamikr | cf2: nice, does it prints those as a command names or just in hex? |
00:25.23 | cr2 | lkcl: but universal/sable have different audio routing. and we don't have PCAP. |
00:26.12 | lkcl | so, cjb: with openmoko on the way, the applications-side is _nearly_ taken care of. the sable is only really suspend/resume-debugging and acx-debugging and GSM-multiplexing (in userspace) away from what you want; cr2 will be able to describe what is needed on the universal, but it involves sound-routing |
00:27.14 | lamikr | lkcl: You have sounds working with sable? |
00:27.16 | cjb | the sable looks more fun than the universal. |
00:28.14 | lkcl | well... it's got GPS. but the sable only has GSM/GPRS, whereas the universal has GSM/GPRS and 3G UTMS |
00:28.49 | lkcl | or, if you _really_ want a challenge, get an htc hermes, instead. |
00:29.15 | lkcl | it has a samsung s3c2442 processor so the battery life on standby is six DAYS instead of 36 HOURS. |
00:29.22 | cr2 | lamikr: it does hex, because i was mostly interested in TI devspecific stuff. but you can take the array from hcidump and add to the program. |
00:29.40 | lkcl | but we haven't even started on it, yet. |
00:30.11 | lkcl | and it's early days for the samsung arm processors. the pxa2xx series, with intel selling its design to maxim or whoever-it-was, is now dead. |
00:30.23 | cr2 | marvell |
00:30.29 | lamikr | cr2: Yeah. Does the hciattach command work for you btw. everytime? In h6300 we have some problem and need usually repeat the command couple of times. |
00:30.51 | lkcl | nobody's buying that's it marvell nobody's buying marvell PXA2xx processors even though they are identical. |
00:31.04 | cr2 | lamikr: yes. but there is a mysterious BT irq. i don't know its purpose. |
00:31.17 | lkcl | the reason is, i believe, because intel _didn't_ sell their royalty-free ARM license, which they acquired when arm were poor and strapped for cash. |
00:31.49 | lamikr | cr2: So hciattach will always work for you? |
00:31.59 | cr2 | intel ? they got it from DEC. strongarm was done by DEC. |
00:32.03 | lkcl | whereas marvell still need an ARM 'distribution' license, which will cost them, as it will push up the price of the processor (for buyers) |
00:32.05 | cr2 | lamikr: yes. |
00:32.36 | lkcl | _license_ - not the arm design. the arm _license_. |
00:32.51 | cr2 | lkcl: they added a h264 decode on chip. |
00:32.59 | lkcl | ahh |
00:33.02 | lamikr | cr2: hciattach newer works for us in the first run. I have not yet investigated it, but I have started to think whether there could be reliability problems in the ttyS0 read and write code. |
00:33.23 | cr2 | lkcl: i'm not a cpu expert, but arm was a shitty design before DEC did strongarm. |
00:33.30 | lkcl | oh. they're an embedded company, aren't they. |
00:34.08 | cr2 | i still have jornada820 with strongarm :) |
00:34.35 | lkcl | hehe. yeh. 600mhz or so in .18 micron when everyone else was using .35 |
00:34.49 | lkcl | oh _that's_ why you have the jornada820.sf.net site. |
00:35.00 | cr2 | lamikr: hmm. with omap you can't just trace the serial port ? |
00:35.02 | lkcl | anyway. ENOUGH! :) |
00:35.06 | cr2 | :) |
00:35.12 | lkcl | i must stop distracting everyone - and myself. |
00:35.58 | lkcl | so, cjb: i look forward to hearing that you got yourself a new toy. either way, i can help out because i have both of them. |
00:36.04 | lamikr | cr2: Propably I can. Need to investigate. |
00:36.14 | lkcl | night folks, and sorry for the noise. |
00:36.21 | cjb | lkcl: sounds good :) |
00:36.47 | cr2 | lamikr: it works for us, but HTC does some additional setup. |
00:36.58 | cjb | It's a shame the openmoko phone doesn't have a keyboard. |
00:37.15 | cr2 | cjb: you don't have any use for UMTS, since you are in the US. |
00:37.44 | cjb | yup. |
00:37.46 | cr2 | openmoko hardware is... hmm... |
00:37.50 | cr2 | a joke. |
00:37.50 | lamikr | bts scripts for h6300 had also some comments for configuring gpio's arn armio's with hci commands. |
00:37.59 | cjb | cr2: heh. how so? |
00:38.12 | cr2 | cjb: compare with the universal :) |
00:38.40 | cr2 | lamikr: TI brf* is a arm920 core. |
00:39.03 | cr2 | it's a computer in itself. with i2c, spi, gpio, etc. |
00:39.41 | BabelOued | good night all :) |
00:39.50 | cr2 | BabelOued: good night. |
00:41.23 | cr2 | cjb: universal and hw6915 are nice devices, but they are expensive too. |
00:42.12 | cjb | so I see. I think I looked at the Sable a couple of years ago. |
00:46.00 | cr2 | years ? |
00:46.09 | cr2 | it's a relatively new phone. |
00:46.11 | cjb | think so, but I might be misremembering/wrong. |
00:46.12 | cjb | ah, ok. |
00:46.31 | cr2 | universal and sable are currently in production. |
00:46.46 | cjb | oh! Are there any US carriers offering them? |
00:47.22 | cr2 | universal is the top of the line device, but i think you've got it already. |
00:47.22 | cr2 | does not make sense for them |
00:47.23 | cr2 | there is no WCDMA2100 network in the US AFAIK. |
00:47.30 | cjb | Ah. |
00:47.44 | cjb | How about the Excalibur? (And has anyone started hacking on it?) |
00:47.57 | cr2 | is it pxa270 ? |
00:48.18 | cr2 | i don't think somebody works on it. |
00:48.41 | cr2 | but if it's pxa, you have very good chances to get a working device really fast. |
00:48.49 | cr2 | and i think it is. |
00:50.46 | cr2 | ok. i'm going to bed too. |
00:50.48 | cr2 | good night. |
00:50.57 | cjb | g'night :) |
02:34.03 | *** join/#htc-linux soze49 (n=soze49@OL180-131.fibertel.com.ar) |
03:08.59 | soze49 | hi Kevin2, r u there ? |
03:57.56 | *** join/#htc-linux scatterp- (n=zsirc@spc2-ashf1-0-0-cust798.asfd.broadband.ntl.com) |
03:58.18 | *** part/#htc-linux soze49 (n=soze49@OL180-131.fibertel.com.ar) |
03:58.26 | *** join/#htc-linux jgw (n=jgw@c-24-22-48-124.hsd1.or.comcast.net) |
03:58.58 | *** part/#htc-linux jgw (n=jgw@c-24-22-48-124.hsd1.or.comcast.net) |
06:48.13 | *** join/#htc-linux goxboxlive (n=goxboxli@gw.hjelcowi.no) |
07:41.12 | *** join/#htc-linux rob_w (n=rob_w@p549B8730.dip0.t-ipconnect.de) |
08:05.23 | *** join/#htc-linux Fossi (n=cmewes@195.177.48.115) |
08:07.31 | *** join/#htc-linux WizMaui (n=WizMaui@62.112.90.231) |
08:43.21 | *** join/#htc-linux rmoravcik (n=rmoravci@pc-3s0zt5w2e4y0vzmhnrzq3a21zqajzfw.users.student.utc.sk) |
08:54.37 | *** join/#htc-linux lkcl_work (n=535411cb@217.147.94.29) |
08:57.33 | *** join/#htc-linux lkcl_work (n=535411cb@217.147.94.29) |
09:03.48 | *** join/#htc-linux lkcl_work (n=535411cb@217.147.94.29) |
09:04.42 | *** join/#htc-linux lkcl_work (n=535411cb@217.147.94.29) |
09:05.31 | *** join/#htc-linux lkcl_work_ (n=535411cb@217.147.94.29) |
09:05.42 | lkcl_work_ | testing testing 123 from cgi irc gateway... |
09:05.55 | lkcl_work_ | o cool it worked. bye! |
09:06.02 | *** join/#htc-linux booba (n=booba@AAnnecy-256-1-55-17.w90-10.abo.wanadoo.fr) |
09:14.11 | goxboxlive | hi lkc_work |
09:16.57 | lkcl | allo goboxlive. just wanted to test the cgi irc gateway so i can go online here, at work. but first i have to _get_ to work :) |
09:56.21 | *** join/#htc-linux lkcl_work (n=c3a976e3@217.147.94.29) |
10:06.19 | *** join/#htc-linux babeloued (n=Fabrice@faim1.itesoft.com) |
10:11.07 | lkcl_work | allo babeloued |
10:14.25 | babeloued | hello :) |
10:15.42 | babeloued | i use http tunnel to use irc :) maybe your idea is better |
10:25.59 | Fossi | http tunnel wins the day ;) |
10:26.42 | Fossi | especially using sshfs through that and streaming music from my disk at home |
10:26.49 | Fossi | poor university ;) |
10:35.58 | *** join/#htc-linux lkcl_work (n=c3a976e3@217.147.94.29) |
11:06.33 | goxboxlive | cr2: What about linux on this device? It has everything, included GPS and UMTS? Fujitsu Siemens Pocket LOOX T830 |
11:08.36 | goxboxlive | http://www.insidepda.de/pocket-pc,Fujitsu-Siemens-Pocket-Loox-T830,technische-daten,167.html |
11:25.33 | cr2 | goxboxlive: it has 240x240 LCD and knowing siemens they will use some arcane CPLD there. |
11:27.12 | cr2 | http://www.insidepda.de/blog/pda-hardware-news/keine-neuen-ipaq-mehr |
11:27.24 | cr2 | hehe. 69xx looks like the last ipaq. |
11:28.12 | goxboxlive | cr2: Ok, so you dont think it will be doabel then in a recent time. |
11:28.30 | cr2 | i wonder how htc will replace universal. |
11:29.04 | cr2 | goxboxlive: we need to have such a device first, and open it. |
11:31.55 | goxboxlive | Well, the rumouers i have been weitten is that it will be the same device, just a little smaller with a bether looking design and with gps included. And i heard that it will me on the market in Q1 2007. Lets hope it has the same hardware as the Universal. |
11:32.12 | goxboxlive | /s/witten/reading/ |
11:33.04 | cr2 | unlikely. what will be the cpu ? |
11:33.32 | cr2 | gps could have been included in the current universal. |
11:33.50 | goxboxlive | yes i know, but they didnt |
11:54.19 | goxboxlive | cr2: The cpu should be samsung as far as i have read. |
12:36.19 | cr2 | i doubt it will be faster. but may have longer standby. |
12:46.44 | *** join/#htc-linux lkcl_work (n=c3a976e3@217.147.94.29) |
12:48.32 | *** join/#htc-linux lkcl_work (n=c3a976e3@217.147.94.29) |
12:49.10 | *** join/#htc-linux lkcl_work (n=c3a976e3@217.147.94.29) |
12:49.14 | *** join/#htc-linux lkcl_work (n=c3a976e3@217.147.94.29) |
12:49.33 | lkcl_work | http://www.howardforums.com/showthread.php?t=1059711 |
12:49.49 | lkcl_work | http://www.whatmobile.net/forum/viewtopic.php?p=400063&highlight=#400063 |
12:55.45 | *** join/#htc-linux lkcl_work (n=c3a976e3@217.147.94.29) |
12:56.29 | *** join/#htc-linux lkcl_work (n=c3a976e3@217.147.94.29) |
12:57.34 | lkcl_work | argh, cgi irc unreliable |
13:01.28 | *** join/#htc-linux booba_ (n=booba@AAnnecy-256-1-57-131.w90-10.abo.wanadoo.fr) |
13:05.40 | babeloued | big CV lkcl :) |
13:07.18 | lkcl_work | cv? oh - yes. sorry! |
13:07.33 | lkcl_work | it's shorter on linkedin |
13:09.10 | babeloued | yes,i see some customer on your CV. Samba TNG is nice project |
13:09.28 | babeloued | I m on linkedin and viaduc too |
13:13.03 | cr2 | lkcl_work: how will they power the usb host on openmoko ? |
13:13.11 | lkcl_work | ahh link me then. |
13:14.46 | lkcl_work | cr2: it's unpowered. you'd have to get a usb hub with 5V external power supply, i imagine. |
13:25.02 | lkcl_work | arg massive delays on cgi irc... |
13:41.33 | *** join/#htc-linux lkcl_work (n=c3a976e3@217.147.94.29) |
13:49.12 | lkcl_work | anyone got an email address for rmoravcik? |
13:49.18 | lkcl_work | rmoravcik: are ya awake, dude? |
13:51.28 | rmoravcik | lkcl_work: hi |
13:56.51 | lkcl_work | oh hello dude. |
13:57.16 | lkcl_work | i wanted to ask you if you could coordinate with pierrox about the g500 patches he's done. |
13:57.29 | lkcl_work | he intends to clean them up some time soon |
13:57.40 | lkcl_work | so i am holding off committing them on his behalf. |
13:58.00 | lkcl_work | i wanted them in so that it would make the hermes port a heck of a lot easier to start |
13:59.05 | rmoravcik | lkcl_work: yesterday rtp from h1940 upload new patches for 2.6.19 |
14:00.27 | rmoravcik | lkcl_work: http://www.rtp-net.org/ipaq/patches/2.6.19/ |
14:02.07 | rmoravcik | it will better try it on hermes or g500 |
14:05.36 | rmoravcik | lkcl_work: rtp want submit these patches to 2.6.20 (mainline) |
14:17.48 | *** join/#htc-linux lkcl_work (n=c3a976e3@217.147.94.29) |
14:40.28 | *** part/#htc-linux goxboxlive (n=goxboxli@gw.hjelcowi.no) |
15:17.33 | *** join/#htc-linux bd2 (n=ufo@89.113.84.151) |
15:23.31 | *** join/#htc-linux lkcl_work (n=c3a976e3@217.147.94.29) |
15:25.54 | *** join/#htc-linux lkcl_work (n=c3a976e3@217.147.94.29) |
15:27.03 | *** join/#htc-linux lkcl_w (n=c3a976e3@217.147.94.29) |
15:27.32 | lkcl_w | i'm so embarrassed, i have to use internet destroyer to connect to cgi irc! |
15:27.57 | booba | erk |
15:28.50 | pof | hi lkcl ... I saw you got an Hermes :D :D |
15:36.45 | *** join/#htc-linux lkcl_w (n=c3a976e3@217.147.94.29) |
15:36.55 | lkcl_w | pof: yep. and cr2 has one as well |
15:37.13 | lkcl_w | argh that didn't work either. i must have screwed up the cgi irc install or something on my server. |
15:37.24 | pof | oh, great news then!! |
15:37.25 | lkcl_w | this isn't really tolerable so i'll go home soon. |
15:38.34 | lkcl_w | i've been encouraging pierrox and rmoravcik to talk to each other, so that the s3c24xx ports join forces |
15:39.18 | pof | i saw there's some support for s3c24xx on latest haret from CVS |
15:39.37 | rmoravcik | lkcl_w: i didn't talk with pierrox yet, only with guys from h1940 port (s3c2410) |
15:40.02 | lkcl_w | oh cool - Kevin2 _did_ integrate that stuff. |
15:40.23 | lkcl_w | i couldn't find my micro sd card adapter so couldn't test it but i have it now. |
15:41.33 | lkcl_w | rmoravcik: he's at work, and busy for a few days. |
15:41.40 | rmoravcik | lkcl_w: i'm using still old haret 0.3.6, but i don't if s3c24xx support in haret isn't broken |
15:42.28 | rmoravcik | lkcl_w: i didn't test latest haret |
15:42.30 | lkcl_w | i'll test this evening and let you know. |
15:42.57 | *** join/#htc-linux jeanseb (n=jeanseb@88.164.32.155) |
15:53.04 | *** join/#htc-linux lkcl_w (n=c3a976e3@217.147.94.29) |
15:53.20 | cr2 | lkcl_w: the hermes arrived!!! |
15:53.40 | cr2 | looks really slick |
15:54.47 | lkcl_w | cr2: yaay! |
15:54.55 | lkcl_w | yeh i thought you might like it. |
15:55.22 | lkcl_w | i think you will like the 6 day battery life, too, compared to the pxa-based devices. |
16:12.16 | rob_w | 6 days .. wow |
16:14.54 | pof | 6 day |
16:15.00 | pof | if you don't enable UMTS |
16:15.17 | pof | otherwise it's just 1 day with data connection always on |
16:15.57 | pof | anyone tried to run the skype linux arm on universal ? |
16:16.24 | pof | it can be extracted from SMC skype phone firmware upgrade :) |
16:18.17 | pof | ftp://asterx.upc.es/soft/skype.linux.arm..EN_17_UK-V1.0.0.10.tar.bz2 <--- download here :P |
16:23.23 | cr2 | pof: does it run over umts data channel ? |
16:24.51 | cr2 | hmm. i think it just needs an ip connection. then you can run it over usb or wifi or bt or IR :) |
16:25.05 | cr2 | not yet umts. |
16:27.48 | pof | cr2: i haven't tried it |
16:45.29 | *** join/#htc-linux psokolovsky (n=psokolov@ip.85.202.124.214.dyn.sub-9.broadband.voliacable.com) |
16:50.24 | cr2 | LOL. the screen protector says i must read wince eula. |
16:51.43 | cr2 | the headset has some weird pseudousb-like connector. |
16:51.53 | *** join/#htc-linux pH5 (n=ph5@e178217059.adsl.alicedsl.de) |
16:52.28 | pof | its standard usb connector |
16:52.39 | pH5 | hey |
16:52.49 | pH5 | lkcl: I'll send you my patches against asoc 0.12.4 |
16:53.20 | pH5 | lkcl: asoc in hh.org cvs has to be upgraded to a more recent version |
16:54.10 | cr2 | pH5: i have changed the ba_audio code to set the source1 CDEX clock. the original code switched the same CONTROL_CX clock twice. |
16:54.40 | cr2 | the blueangel_pcmcia clock init is also buggy. taken from hx4700. which is also buggy ;) |
16:55.55 | cr2 | oh. hermes has 2 cams ? |
16:56.18 | pH5 | hi cr2 |
16:56.20 | pH5 | yes it has |
16:56.26 | cr2 | the small one has very bad optics ;) |
16:56.33 | cr2 | hi pH5 |
16:56.39 | pH5 | yeah, I guessed that |
16:57.35 | cr2 | what is this "slashed" usb ? |
16:57.44 | *** join/#htc-linux goxboxlive (n=goxboxli@9.80-202-160.nextgentel.com) |
16:57.57 | pH5 | slashed usb? |
16:58.14 | cr2 | doublesided, with a cut corner. |
16:58.29 | cr2 | it's not miniUSB a/B |
16:59.05 | pH5 | ah, on the hermes? nonstandard connector. miniusb + headset pins |
16:59.16 | pH5 | or something like it |
16:59.46 | pof | cr2: see wiki Hermes_FAQs |
16:59.50 | cr2 | ok. |
16:59.51 | pof | Is the MiniUSB connector standard? |
16:59.57 | pof | The port is NOT proprietary. It is a standard mini-USB, and takes standard mini-USB cables. |
16:59.57 | pof | They tweaked the socket a little bit to go along with their new proprietary headset. |
17:00.07 | cr2 | it charges from my universal AC. good. |
17:00.19 | pH5 | exactly. it just has additional pins on the other side of the connector |
17:00.49 | cr2 | miniUSB is of course standard. there are 2 connectors/sockets. A (host) and B(slave) |
17:01.22 | cr2 | 2 megapix. what camera is that ? |
17:01.29 | pH5 | main one |
17:01.48 | cr2 | i mean not ov9650. |
17:02.02 | pof | http://www.tracyandmatt.co.uk/blogs/index.php/2006/09/10/htc_hermes_usb_connector_pin_config |
17:02.36 | pH5 | pof: cool |
17:03.12 | cr2 | pof: switch pin ? is there a button on the headphone ? |
17:03.15 | cr2 | it does not work on universal. |
17:03.15 | pof | yes it is |
17:03.37 | cr2 | ok. one more gpio to look for :) |
17:03.37 | pof | audio volume and hangup / answer button is on the headphne |
17:03.41 | pH5 | I have a button on the magician, too. |
17:03.50 | pH5 | But here everything is just on pxa gpios :-P |
17:03.56 | cr2 | but it works on blueangel. |
17:04.47 | goxboxlive | cr2: I have allready told you about the button on the headphones: I tiold you that had to be the missing GPIO |
17:05.00 | cr2 | this eusb is very loose. not a great machanical design. |
17:05.09 | cr2 | goxboxlive: there is no gpio ;) |
17:05.35 | pH5 | cr2: but it's much better than the 2.5mm shit I have to deal with |
17:06.26 | cr2 | it will not last long. |
17:06.33 | cr2 | so only universal has 3.5" ? |
17:07.02 | *** join/#htc-linux WizMaui (n=WizMaui@62.112.90.231) |
17:07.05 | goxboxlive | cr2: Ok, but the button is used for two things: 1) To answer a incomming call 2) If you press it for a while (3 sec. or so) you activate voice speed call. |
17:07.40 | cr2 | goxboxlive: i didn't see any gpio change on this button press. |
17:08.21 | cr2 | uuh. microSD. |
17:09.03 | cr2 | pof: how do i open the case to insert the battery ? |
17:09.24 | cr2 | ok, found. |
17:11.25 | cr2 | the fon is so tiny :) |
17:11.51 | cr2 | s/fon/font/ |
17:23.34 | lkcl | ph5 - are you philip zabel? |
17:23.49 | pH5 | <PROTECTED> |
17:23.51 | lkcl | you are, aren't you? :) |
17:23.53 | lkcl | o yeh |
17:23.54 | pH5 | yes :) |
17:23.59 | lkcl | cool |
17:24.16 | lkcl | liam says you have some asoc magician code kicking around? |
17:24.27 | lkcl | if so is it around somewhere i can play with? |
17:24.35 | lkcl | i neeeeeeed to get blueangel sooooooound |
17:24.53 | lkcl | i neeeEeeEeeeeEEEd it |
17:24.55 | pH5 | I'll send you my patches against asoc 0.12.4 |
17:25.15 | lkcl | uhkaaay. gotta get my coding fiiiix for the niiight |
17:25.17 | pH5 | could be that it's totally broken again after liam's latest changes, though |
17:25.25 | lkcl | o well i don't care ha ha |
17:25.46 | lkcl | well it will give me something to do and save you some time |
17:32.18 | pH5 | lkcl: mail sent, asoc 0.12.4 is at http://opensource.wolfsonmicro.com/~lg/asoc/asoc-v0.12.4.patch |
17:34.31 | pH5 | lkcl: so please port the ba driver and fix uda1380 so I get some sound, too :) |
17:34.57 | babeloued | lkcl, pH5 : me too :) |
17:36.31 | cr2 | pH5: the asoc is very wolfsonmicro oriented stuff. why don't they target ak4641, which is in active use and production ? |
17:36.58 | pH5 | probably because they care most about their own codecs? |
17:37.12 | pH5 | I bet all important wm chips work currently. |
17:37.12 | cr2 | <PROTECTED> |
17:37.45 | cr2 | still the have ak4531 |
17:37.54 | cr2 | s/the/they/ |
17:41.02 | cr2 | the wifi on hermes is not so sensitive as on ba/uni/sable. |
17:42.02 | *** part/#htc-linux babeloued (n=Fabrice@faim1.itesoft.com) |
17:53.34 | goxboxlive | hmm gomunicator failes cause of some 'dbus-marshall-recursive security policies. |
17:57.48 | cr2 | lol |
17:58.28 | cr2 | i want a ncurses version. |
18:05.45 | *** join/#htc-linux rob_w (n=bob@p85.212.130.78.tisdip.tiscali.de) |
18:10.55 | *** join/#htc-linux rob_w (n=bob@p85.212.130.78.tisdip.tiscali.de) |
18:11.55 | goxboxlive | cr2: What is this openmoko disussion going on. Will we be able to run openmoko in example on angstrom distro? Or is the openmoko the distro with the GUI. If so, is there a way we can use the GUI on another distro than openmoko? |
18:15.36 | *** join/#htc-linux BabelOued (n=Fabrice@82.238.28.28) |
18:50.53 | *** join/#htc-linux skodde (n=skodde@unaffiliated/skodde) |
19:02.40 | goxboxlive | cr2: Have you looked at this driver? Does it fit our system? USB driver for GSM and CDMA modems (USB_SERIAL_OPTION) |
19:04.25 | rob_w | ~ openmoko is a opensource initiative from FIC |
19:04.27 | apt | ...but openmoko is already something else... |
19:04.37 | rob_w | ~ what is openmoko |
19:04.39 | apt | rob_w: what are you talking about? |
19:05.34 | rob_w | ~openmoko? |
19:05.35 | apt | i heard openmoko is a joint venture of FIC and the open source community creating a complete open source phone platform stack. see http://openmoko.com |
19:05.58 | rob_w | well where "phone platform stack" is a bit blurry here |
19:06.50 | rob_w | the only big thing i think is a successor to gtk , been hidden for years and totally open source |
19:14.39 | cr2 | goxboxlive: it's a distro with it's own gtk-based UI. right now it is vaporware. useful parts are the phone app and the USB mux driver. |
19:15.10 | goxboxlive | ok |
19:19.29 | *** join/#htc-linux Fossi (n=Fossi@e176097134.adsl.alicedsl.de) |
19:27.27 | goxboxlive | cr2: Are you around? |
19:27.43 | rob_w | vaporware eh ? |
19:27.45 | goxboxlive | Stop playing with that hermes phone and get back to the universal :-) |
19:28.02 | rob_w | cr2, so they do use gtk , interesting |
19:30.50 | cr2 | rob_w: if one can't show anything except press releases, i can it vaporware :) |
19:31.23 | rob_w | so that would have been my only logical meaning, nice |
19:31.43 | cr2 | goxboxlive: universal is so easy :) there are no tools for hermes actually. |
19:31.50 | goxboxlive | cr2: OE developers has removed detect-stylus from the htcuniversal. |
19:31.54 | rob_w | in the end its all standard beside a matchbox-theme |
19:32.10 | goxboxlive | They says that we shall begin to use /dev/touchscreen0 |
19:32.24 | cr2 | how ? |
19:32.28 | goxboxlive | And also that udev will make that rule. |
19:32.35 | goxboxlive | or make a symlink |
19:32.38 | rob_w | ln -s maybe |
19:32.53 | cr2 | i plug the modules in random order, and your symlinks are dead. |
19:33.06 | rob_w | <PROTECTED> |
19:33.28 | cr2 | is there a unique id for an input device ? |
19:33.33 | cr2 | rob_w: sure :) |
19:33.44 | goxboxlive | But, in opie, there is a link to detect-stylus, should i just make a hardlink to touchscreen0? |
19:33.52 | rob_w | i bet evdev bring some idtype things under /sys |
19:34.07 | cr2 | in opie ? |
19:34.12 | goxboxlive | Or what about ts_calibrate, it doesnt work cause it miss ts_lib or something. |
19:34.26 | goxboxlive | I'll try out a new fresh image and see whats happends |
19:34.35 | goxboxlive | Psokolovsky: Are you around? |
19:34.36 | cr2 | in /etc/init.d/opie , but it's shell script. |
19:34.45 | goxboxlive | yes i know |
19:34.53 | cr2 | ts_calibrate ? |
19:34.55 | rob_w | well tslib does always check /dev/touchscreen0 but also respects TSLIB_TSDEVICE |
19:35.26 | goxboxlive | In opie it look like this: export QWS_MOUSE_PROTO=TPanel:`detect-stylus --device` |
19:35.30 | cr2 | i think opie has a bulitin (btw, not affine) calibration routines ? |
19:36.04 | cr2 | goxboxlive: if you want to hardcode /dev/touchscreen0 , then you can just change the QWS_MOUSE_PROTO= |
19:36.05 | rob_w | goxboxlive, it is the equivalent to the export for TSLIB |
19:36.10 | goxboxlive | yes it has, but i have tried once before with this export QWS_MOUSE_PROTO=TPanel:/dev/input/event0 and it didnt work |
19:37.01 | cr2 | goxboxlive: but `detect-stylus --device` is /dev/input/event0 :) |
19:37.31 | goxboxlive | hmm but why didnt it work ? |
19:37.37 | cr2 | i think the input device must have a unique id, and it can be hten hardcoded. |
19:37.47 | goxboxlive | should i use ` |
19:37.50 | cr2 | goxboxlive: then you did something wrong :) |
19:37.55 | cr2 | no. |
19:37.56 | goxboxlive | :-) that might be |
19:38.09 | cr2 | it's a normal shell substitution. |
19:38.20 | goxboxlive | well i give it a try, with event0 |
19:38.25 | cr2 | exec this command, and put the text string at its place. |
19:38.44 | rob_w | goxboxlive, but make sure to load it first before any other evdev device ;-? |
19:39.14 | cr2 | rob_w: ? evdev is hardcoded. |
19:39.41 | rob_w | ?? |
19:39.44 | rob_w | since when ? |
19:43.01 | cr2 | i did it to avoid exactly the same PITA like on .12 ba :) |
19:49.04 | rob_w | u again , grr |
19:53.27 | goxboxlive | Well, udev didnt create touchscreen0, and i had to also edit tslib.sh to point to event0. Bit, now, at lest it's working. |
19:55.31 | cr2 | ok. |
19:56.54 | psokolovsky | goxboxlive, hi! |
19:57.02 | goxboxlive | hi there |
19:57.08 | cr2 | goxboxlive: the audio switching system has at lead 2 in-kernel and 12 static variables (!) to control the audio state. |
19:57.09 | goxboxlive | one question |
19:58.07 | goxboxlive | psokolovsky: After oe.dev removed detect-stylus from tslib. Is your udev creating a touchscreen0 without doing anything to udev rules? |
19:58.41 | psokolovsky | goxboxlive, I din't check that for some time. but the rule should be there by default. |
19:58.44 | goxboxlive | cr2: so that will say that it is more complex that forst thought. |
19:58.55 | goxboxlive | ok thx psok. |
19:59.08 | rob_w | why does dbus always needs extra attention these days when building on oe ? |
19:59.19 | cr2 | goxboxlive: yes, but not all of them are used. |
19:59.41 | cr2 | psokolovsky: what is the purpose of the .id in platform_device ? |
19:59.54 | goxboxlive | cr2: Take a look at this line. Do you see something wrong wich causes udev to not make a touchscreen0? |
20:00.05 | goxboxlive | SUBSYSTEM=="input", KERNEL=="event[0-9]*", SYSFS{modalias}=="input:*-e0,1*,3,*a0,1,*18,*", SYMLINK+="input/touchscreen0" |
20:01.30 | cr2 | goxboxlive: to be honest i don't see a point for using udev here. |
20:02.07 | cr2 | maybe a match of input->name ? |
20:02.28 | *** part/#htc-linux baud123 (n=baud@ver78-1-82-240-29-152.fbx.proxad.net) |
20:02.40 | *** join/#htc-linux g3gg0_ (n=g3gg0@ppp-62-245-162-80.dynamic.mnet-online.de) |
20:03.22 | psokolovsky | cr2, if there're more than 1 device of of the same type, to identify them |
20:06.12 | cr2 | ok. |
20:19.24 | goxboxlive | well that's it, detect stylus is now removed, and touchscreen is working with linking to event0 |
20:22.03 | psokolovsky | goxboxlive, more exactly, TS is working to linking by *whatever* offers touchscreen event input. so, if you'll have it on event10, it still will work. |
20:23.00 | goxboxlive | yes, i have done it that way. linking to event0 |
20:25.01 | cr2 | psokolovsky: what will happen if i'll connect the bt mouse ? |
20:25.48 | cr2 | we also need a solution for merging 3 keyboards together, after the gpio-keys, asic3-keys and pxa27x-keyboard split. |
20:26.12 | psokolovsky | cr2, happens with what? |
20:27.22 | psokolovsky | cr2, I said that every time, say it again: there is NO explicit solution needed for that, it is handled automagically. At least, that's how it works on h4000 & hx4700. If youhave any issues with that, give me more info, please. all info ;-) |
20:32.17 | cr2 | will the bt mouse be found with this udev rule ? |
20:33.08 | cr2 | the keyboard problem is pretty obvious. imho, there is no way to tell opie to sample from N keyboards simultaneously. |
20:33.48 | *** join/#htc-linux psokolovsky_ (n=psokolov@233.usernat.ip.net.ua) |
20:41.48 | *** join/#htc-linux psokolovsky_ (n=psokolov@233.usernat.ip.net.ua) |
20:58.09 | *** join/#htc-linux asylume2 (n=insanity@196.211.28.91) |
21:04.33 | goxboxlive | cr2: Are you around |
21:06.26 | goxboxlive | The OE guys wouldnt fix the tslib for us. They said that we had to fix the hardcoding of evdev. Can we fix it? Why do we hardcode it? |
21:14.46 | *** join/#htc-linux rejo1 (n=rejon@adsl-75-5-124-98.dsl.pltn13.sbcglobal.net) |
21:14.57 | *** part/#htc-linux rejo1 (n=rejon@adsl-75-5-124-98.dsl.pltn13.sbcglobal.net) |
21:32.51 | cr2 | don't ask me about such deep userspace issues :) |
21:32.59 | goxboxlive | :-) |
21:33.24 | cr2 | i'll better look at the audio stuff today. |
21:33.34 | goxboxlive | yes thx |
21:34.04 | cr2 | we need a working phone. the rest is not so important. |
21:34.57 | goxboxlive | Thats right, a working phone and suspend/resume working, then it will be a usefull device |
21:35.24 | cr2 | i think the pxa27x_keyboard is not a platform device. |
21:35.40 | goxboxlive | ok, what do we do then? |
21:36.22 | cr2 | i'll try to correct it now. |
21:41.01 | cr2 | compiling. |
21:41.48 | cr2 | goxboxlive: what about btsco ? |
21:47.09 | goxboxlive | well , i have tried it, also build a modules btcsc and installed it, but the universal wouldt connect to the BT headset. Have to look further in to it. Maybe creating rfcomm and then try. |
21:48.02 | cr2 | done. |
21:48.08 | cr2 | check the CVS. |
21:49.24 | goxboxlive | what is done? i saw that you updated the keyboard driver. But what is working? |
21:50.31 | cr2 | it is platform_device not, otherwise it will not resume. |
21:50.45 | cr2 | and will crash. |
21:51.19 | cr2 | so try to suspend and resume. |
21:51.34 | goxboxlive | ok, so the keybaord resumes now? And the power button is also working? |
21:53.03 | cr2 | check. |
21:53.05 | cr2 | the keyboard will not resume, i think. but at least is should not crash. |
21:53.37 | goxboxlive | ok, i am compiling now. |
21:54.04 | lkcl | pH5: i've committed your stuff. if it makes your life awkward for further stuff, i take responsibility and fix it by doing cvs merges for you ok? |
21:55.34 | cr2 | lkcl: i'm not impressed with the hermes wifi performance. |
21:55.46 | lkcl | good god, you got wifi? |
21:55.57 | cr2 | not in linux :) |
21:56.03 | cr2 | for haret. |
21:56.06 | lkcl | have you got the hermes firmware? |
21:56.24 | cr2 | what do you mean ? |
21:56.51 | cr2 | it 's a good question how much changes to acx100 are needed to make it work in linux. |
21:57.32 | cr2 | the biggest problem is that s3c24xx does not have "wi". |
21:57.51 | cr2 | and 0.4.4 does not have "dump gpio", wgpio , etc. |
21:57.59 | cr2 | lsmod and ps work. |
21:58.22 | cr2 | i'm already thinking about evil uses for lsmod :) |
21:58.51 | lkcl | the firmware for hermes is different, remember? |
21:58.58 | lkcl | hermes, hermes.... o oops wrong device ha ha |
21:59.05 | lkcl | i thought you meant sable |
21:59.32 | lkcl | oh - you got a boot already? |
22:01.39 | cr2 | :) no. i need to write some "dump" code first. let's hope Kevin2 will merge the wgpio code. |
22:02.03 | cr2 | lkcl: btw, what about registering the mtype for greenphone ? |
22:02.13 | lkcl | oh |
22:02.16 | lkcl | yeh |
22:05.25 | *** join/#htc-linux ljp (n=lpotter@203.94.178.46) |
22:06.19 | lkcl | ljp: did trolltech submit a mach id for greenphone? |
22:07.18 | ljp | dunno. maybe yuhuatel |
22:07.59 | cr2 | ljp: can you tell us which sound and bluetooth chipset are used ? |
22:08.25 | lkcl | ok |
22:11.20 | cr2 | this one ? http://www.yuhuatel.com/english/omega.php |
22:11.53 | BabelOued | lkcl : where is tsc2200.h ? |
22:11.59 | lkcl | uhhhnn.... ermmm :) |
22:12.02 | ljp | thatd be it |
22:12.40 | goxboxlive | cr2: The keybaord has stopped working, even before suspending. The /dev/input/even1 is reacting when doing 'cat', but it is not resumeing when it's hit. |
22:12.50 | ljp | sound is WM 9713 , bt is stlc2500c |
22:13.07 | cr2 | ljp: thanks. |
22:15.14 | lkcl | o great - wm 9713 is a wolfson codec _excellent_ |
22:15.14 | lkcl | wolfson like asoc, lots |
22:15.37 | lkcl | ok committed BabelOued |
22:15.51 | BabelOued | lkcl: thanks :) |
22:15.56 | lkcl | sorry! |
22:16.24 | cr2 | ./sound/soc/codecs/wm9713.c |
22:26.46 | cr2 | lkcl: http://72.14.221.104/search?q=cache:mNcI8Y1JzDUJ:blackfin.uclinux.org/pm/task.php%3Ffunc%3Ddetailtask%26project_task_id%3D393%26group_id%3D17%26group_project_id%3D74+stlc2500c&hl=de&gl=de&ct=clnk&cd=6&ie=UTF-8 |
22:28.06 | cr2 | 1) Validate Bluetooth HCI over UART with STLC2500C chipset. |
22:28.13 | cr2 | 2) Add driver support for PCM/I2S Interface |
22:29.02 | cr2 | Start Date: 2007-01-12 End Date: 2007-01-17 |
22:30.59 | lkcl | ack |
22:31.15 | cr2 | Sony Ericsson K800i uses STMicroelectronics (STLC2500C) |
22:31.21 | BabelOued | lkcl : blueangel_tsc2200.c:29: error: 'GPIO_NR_BLUEANGEL_TSC2200_IRQ_N' undeclared |
22:31.32 | BabelOued | missing some #define |
22:31.35 | lkcl | BableOued: oops. |
22:31.36 | lkcl | 1sec... |
22:32.23 | lkcl | yeh it is, isn't it. |
22:32.25 | lkcl | no idea why. |
22:33.14 | lkcl | 1sec... |
22:33.40 | lkcl | include/asm-arm/arch-pxa/htcblueangel-gpio.h |
22:35.21 | cr2 | good night. |
22:35.35 | BabelOued | good night cr2 |
22:37.54 | *** join/#htc-linux florian (n=fuchs@87-193-10-210.ipool.celox.de) |
22:38.22 | BabelOued | lkcl : i have tsc2200_KB_N http://handhelds.org/cgi-bin/cvsweb.cgi/linux/kernel26/include/asm-arm/arch-pxa/htcblueangel-gpio.h?only_with_tag=MAIN |
22:38.27 | florian | re |
22:38.39 | BabelOued | it is in MAIN branch ? |
22:38.56 | lkcl | yep |
22:40.04 | *** join/#htc-linux psokolovsky (n=psokolov@ip.85.202.124.214.dyn.sub-9.broadband.voliacable.com) |
22:40.59 | BabelOued | missing two line |
22:41.03 | BabelOued | #define GPIO_NR_BLUEANGEL_TSC2200_IRQ_N 16 |
22:41.10 | BabelOued | #define IRQ_NR_BLUEANGEL_TSC2200_TS IRQ_GPIO(GPIO_NR_BLUEANGEL_TSC2200_IRQ_N) |
22:44.06 | lkcl | yeh i addeed them |
22:48.21 | BabelOued | ok, i ve another compile problem no.. :( |
23:03.01 | BabelOued | lkcl: it compile for you ? |
23:09.57 | BabelOued | lkcl ; i found difference in tsc2200.h |
23:10.37 | BabelOued | maybe the new include line nssp_tsc2200.h |
23:16.13 | *** join/#htc-linux Ralith (n=ralith@66.150.12.86) |
23:21.45 | BabelOued | ok now it compile :) |
23:27.52 | *** part/#htc-linux scatterp- (n=zsirc@spc2-ashf1-0-0-cust798.asfd.broadband.ntl.com) |
23:59.07 | psokolovsky | Kevin2, Hi, online? |
23:59.31 | psokolovsky | Kevin2, would like to discuss ways to provide haret snapshots for download |