irclog2html for #htc-linux on 20070107

00:00.05kimuSan^cr2: did you disable the serial debug ?
00:00.50cr2kimuSan^: yes.
00:01.56kimuSan^ok
00:02.14cr2well, then we need to do more debugging.
00:04.10kimuSan^sure. theres always another day tomorrow :-)
00:04.26pofKevin2: it is a modification to current CVS haret code (pulled yesterday)
00:04.38pofin order to load a Hermes SPL instead of a linux kernel
00:05.09pofwith the new code, i can copy SPL correctly into phy 0x300000000
00:05.22pofbut it hangs when I jump into it
00:06.03Kevin2pof: Are you sure the SPL should be at 0x300000000 and not some other address?
00:06.38cr2pof: where does your SP point to at jump time ?
00:06.40pofi guess it is the correct
00:07.02pofcr2: how do i know that ?
00:07.33cr2do you know where the SPL puts its stack ?
00:07.41Kevin2pof: Do you get the red,blue,black lines?
00:08.09pofKevin2: yes, i can see the red/blue lines on screen
00:08.33kimuSan^Kevin2: what does each of the codes indicate?
00:08.41pofnot the black
00:08.53Kevin2kimuSan^: Checkout haret cvs and read the comments in linboot.cpp.
00:09.22kimuSan^ahh that where it is...
00:09.25pofcr2: i decoded some parts with ida, don't have them here right now :(
00:09.56Kevin2pof: Oh, you commented out the black line.  My guess is it jumps to that address but the code there is not setup properly.
00:10.31pofthe bootloader expects to be executed from virtual address 0x8c080000
00:10.41*** join/#htc-linux rejon (n=rejon@c-67-174-202-239.hsd1.ca.comcast.net)
00:10.42pofso maybe I should try to load it at 0x30080000
00:10.52*** part/#htc-linux rejon (n=rejon@c-67-174-202-239.hsd1.ca.comcast.net)
00:11.10cr2pof: really ? it expects the mmu on ??
00:11.48cr2it's not the case on universal.
00:11.52pofno, I  have to copy the mem to right phisycal offset AFTER unplug the MMU's
00:12.07Kevin2kimuSan^: cvs.handhelds.org:/cvs/haret
00:12.16cr2btw, what is your IPL doing ?
00:12.39pofi'm not sure... there is some interaction between IPL / SPL
00:12.51kimuSan^Kevin2: I have it...just looked through the code briefly but did not find it back then.
00:12.54cr2hmm.
00:12.55pofbut i don't have the disassembled files here... i have ida at work :(
00:13.03cr2ok.
00:13.47cr2pof: it sounds very strange. very different from pxa wince ipl+spl
00:14.03Kevin2kimuSan^: If you have an old checkout, you may want to do a "cvs up".  There is a big comment section which describes the colors at the end.
00:14.12pofwhat does on PXA ?
00:14.49cr2the IPL does some minimal setup and then decides if it's wakeup or cold boot.
00:15.02pofit's the same here then
00:15.08cr2if it's wakeup, it jumps to 0xa0040000
00:15.21cr2"wince kernel"
00:15.42cr2if not, then it loads 256K to 0xa0000000-0xa0040000
00:15.57cr2this all happens with mmu off.
00:16.11cr2IPL is not using the stack.
00:16.27cr2only the cpu registers.
00:16.43cr2well, it's not like shitty x86.
00:17.00kimuSan^Kevin2: just did.
00:17.09pofsome time ago I disassembled ipl and could follow most part of it with the samsung manual
00:17.09cr2256K come from G3 DoC.
00:17.15kimuSan^my cegcc installation seems broken though
00:17.20pofbut i don't have this stuff here, shit!
00:17.28kimuSan^wasn't the a complete toolchain package somewhere?
00:17.35cr2pof: it's good. can be very helpful for linux wakeup :D
00:17.57cr2kimuSan^: no.
00:18.09kimuSan^ok
00:18.15pofi'll try to put everything together on monday night
00:18.48cr2ok.
00:19.14cr2256K is the SPL, forgot to mention.
00:20.37pofyes :)
00:20.37pofi removed the 1st 4096 bytes of it, I remember there is a big jump at the begining
00:20.38cr2so it's pretty straitforward.
00:20.45cr2don't you have some sig in between ?
00:21.11cr2but you may break the references if you just trim the code ?
00:21.13pofdon't knwow, the SPL is much harder to follow and I'm not very expericed with ASM
00:21.37pofoh
00:21.50pofprobably this makes sense, so I have to copy it full, and jump later?
00:21.58cr2spl is _much_ better than wince kernel ;)
00:22.05cr2yes.
00:22.40kimuSan^gaa im too tired tonight. I think I will head to bed. thanks for the try cr2. nite all
00:23.05cr2following something in the wince kernel is really ugly.
00:23.15cr2kimuSan^: good night.
00:24.25cr2it seems to me that porting to samsung is slow just because of zero experience with this cpu. it's not that many settings we need to dump from wince.
00:26.03kimuSan^<PROTECTED>
00:28.28goxboxliveI didnt know that this was working: http://pastebin.ca/308822
00:33.53cr2no much more hype from openmoko. everybody is silent now.
00:36.33*** join/#htc-linux hlbot (n=adm@iclem.net)
00:37.45goxboxliveArent they going to realease it this month?
00:40.11*** join/#htc-linux lkcl (n=lkcl@90.194.116.202)
00:41.11goxboxliveso it's baby feeding now i guess lkcl ? Am i right?
00:41.17lkclhe he
00:41.20goxboxlive:-)
00:41.24lkclwe just had 60 guests, they just all left
00:41.35goxboxlivewow
00:41.44lkclone of them wandering down the middle of the road.
00:41.48goxboxlivethats alot of people
00:41.52lkclit's a big house
00:42.17goxboxliveYes i guess
00:42.25lkclabout 15 in the kitchen.
00:42.59lkcli'm supposed to be going back to holland tomorrow
00:43.03goxboxlivewell in my aparrtment we have just room for us 4 who lives here
00:43.18goxboxlivewhat are you doing in holland? Working?
00:43.22lkclyes
00:43.32lkclsqueeeeeeze :)
00:43.39lkclpop
00:44.12lkclexcept my motherinlaw isn't well
00:44.16goxboxlive:-) We are about to buy us a house, but it is so damn expensive.
00:44.42lkclgoboxlive: are you a spending-kind-of-person or a saving- ?
00:44.53goxboxlivespending
00:45.04lkclah then get a repayment mortgage.
00:45.40cr2lkcl: hi
00:46.02cr2i have added the setup for trinity in CVS.
00:46.04lkclif you trust yourself with money, get an interest-only mortage.  never in your life not in 40 years are you _actually_ going to have a house that doesn't have a mortgage on it.
00:46.07lkclcr2: great!
00:46.14lkclwhich one's trinity?
00:46.23goxboxliveof course lkcl, uo here in Norway you arent abel to even buy an appartment  without a loan.
00:46.37lkclcr2: ah, ha ha ha
00:46.49cr2trinity is a keyboardless hermes
00:46.53lkclyou haven't got one of those already, have you, cr2?
00:47.22cr2no. i have enough of them already :)
00:47.37lkclwhat - you have enough trinity phones?
00:47.44lkclyou can never have enough htc phones.
00:47.53cr2goxboxlive wants to get one for 1300€
00:48.07cr2wallaby. that's a funny phone.
00:48.38lkcli found these people in romania who are willing to sell me one, now that they're part of the EU.
00:48.48lkclthey wouldn't take my money before 3rd jan.
00:49.12lkclok folks i gotta get back to my remaining guests
00:49.14lkclttyl
00:51.34goxboxlivewell i havent bought a phone for over about one and a half year so i guess it's about time to get a new one. Though i have to see what it will cost here up north.
00:55.11cr2goxboxlive: check it with 7mobile.de if it's a reasonable price.
00:55.43goxboxliveBut isent that with contract?
00:56.31cr210€month=+240€, and you can talk for these 10€/month.
00:56.38cr2i think it's reasonable.
00:59.11goxboxliveBut they wouldnt sell phone if you are not a Deutch right? In norway you cant buy a phone with contract if youre not a norwgian citizent
01:01.57goxboxlivewell its getting late. g'night
01:03.07BabelOcr2: what is the problem in this code ? http://pastebin.ca/308872
01:04.11cr2<PROTECTED>
01:04.11cr2<PROTECTED>
01:04.24cr2what if ret!=nbytes?
01:07.48cr2good night.
01:08.02BabelOcr2: thanks good night
01:19.46*** join/#htc-linux Hawk|- (n=Hawk@p57A54822.dip0.t-ipconnect.de)
01:41.19lkclBabel0: look in samba tng lib/util_sock.c and also in xanadux.sf.net proxyd code for same util_sock.c - there's some code that does socket reading etc. etc.
01:41.30lkclread doesn't _have_ to return exactly the number of bytes you request
01:42.10BabelOlkcl: ok, because libgnokii doesn't parse well
01:58.55*** join/#htc-linux Hawk||- (n=Hawk@p57A51C89.dip0.t-ipconnect.de)
02:24.03*** join/#htc-linux maniacxs (n=maniacxs@pD95468AD.dip0.t-ipconnect.de)
03:05.18*** join/#htc-linux phongred (n=phongred@203.113.144.83)
03:05.25phongredhi all
03:06.44phongredyou can help me setup linux with HTC BlueAngel Wm2k5,
03:15.35rpodgornywhat's the problem?
03:22.27pofphongred: i think on BA you can only boot linux when run on WM2003
03:22.55pofon WM5 it never worked for me, although you may try with newer haret version from CVS
03:23.37pofhttp://handhelds.org/~koconnor/haret/
03:26.14rpodgornybooting worked for me on wm5, haven't tried much else thou...
03:28.54pofok, that may have changed then... i don't use BA since long time :)
03:32.57rpodgornyjust follow the procedure from hh.org wiki and it should work
03:32.58rpodgornyhttp://www.handhelds.org/moin/moin.cgi/BlueAngel
03:39.16phongredthanks rpodgorny
03:41.42phongredyou can help me setup linux?
03:42.00phongredsorry  my englidh very bad lol!
03:42.06phongredenglish
03:42.17phongredi'm using Wm2k5
03:44.09rpodgornyjust follow the procedure carefully and it will be fine
03:44.54phongredyou can help me begin to end.lol
03:44.56phongredthanks
03:45.49phongredrecommand storge card ?mb
03:45.59rpodgornyi have 1gb
03:46.06phongredlol
03:46.09phongredoh no
03:46.30phongredminimum?
03:47.31rpodgornywell, 128 or 256 should be good enough
03:47.39phongredyes thank
03:47.47phongrednow u can help me?
03:48.25phongredwhere file linux down load
03:48.44rpodgornyit's all written there
03:49.22phongredyou can send me link download file linux for Blueangel WM5
03:49.43rpodgornygotta go to bed, good nite all...
03:50.06phongredhix
03:51.43phongredhelp me
04:17.15phongredhixhix
04:17.16phongredhelp me
05:12.54*** join/#htc-linux BobOfDoom (i=ato@2002:d20b:925d:0:0:0:0:1)
05:19.23VisNphongred, use cvs
05:22.48VisNphongred, http://www.handhelds.org/moin/moin.cgi/GettingHandheldKernels
05:24.57VisNor from ftp ftp://ftp.handhelds.org/misc/linux/kernel/
05:26.19VisNbut kernel's from ftp too old
05:27.49VisNfrom cvs often is broken...
05:51.59phongredi'm here
05:54.12phongredftp://ftp.handhelds.org/misc/linux/kernel/  <== i dont know file of BlueAngel
05:58.25VisNoh, probably u are looking for distributions
05:59.10phongredu can help me
05:59.13phongredlol
05:59.14phongredthanks
05:59.33phongredi'm using BlueAngel WM5
06:01.01phongredharet.exe   <== cant run with WM2005
06:01.12phongredonly run with 2003wm
06:02.14iggyhave you checked the url in the /topic that says "please check"?
06:05.02phongredhic
06:08.26phongredi open this link  http://www.handhelds.org/moin/moin.cgi/BlueAngel  , but This Haret.exe does NOT WORK with WM2005 on Blueangel.
06:09.23VisNhttp://handhelds.org/moin/moin.cgi/BlueAngel
06:09.34phongredyes
06:09.44phongred<PROTECTED>
06:10.04phongredu can help me, because i'm using BA WM2005
06:10.08VisNI see
06:10.12phongredlol
06:10.26phongredi'm from vietnam
06:10.39phongredmy english very bad
06:10.41phongredlol
06:11.11iggyThis Haret.exe does NOT WORK with WM2005 on Blueangel. You will only See the Indicator after it is full nothing will happen. Try to add the "mem=128M" to the kernel boot line.
06:11.18iggyfrom that page
06:11.52phongred<PROTECTED>
06:12.00phongredi dont know
06:12.23phongred<PROTECTED>
06:12.26phongred?
06:13.57VisNphongred, I hope you understand what you'll lose all your data from RAM? contacts, programs...
06:14.49phongredyes, i see
06:15.06phongredstep 1: backup data
06:17.55phongredbut i dont know  "mem=128M"
06:18.07phongredi dont undestand
06:20.46phongredhix
06:20.50VisNlook in startup.txt
06:21.40phongredset KERNEL "zImage-2.6.12"
06:21.41phongredset MTYPE 766
06:21.41phongredset INITRD "initrd-2.6.12-hh2.gz"
06:21.41phongredset CMDLINE "root=/dev/ram0 init=/linuxrc ramdisk_size=14336 keepinitrd console=tty0 boot_mmc=y"
06:21.41phongredbootlinux
06:22.10phongredchange to
06:22.10phongredset KERNEL "zImage-2.6.12-hh2"
06:22.10phongredset MTYPE 766
06:22.10phongredset INITRD "initrd-2.6.12-hh2.gz"
06:22.10phongredset CMDLINE "root=/dev/ram0 init=/linuxrc ramdisk_size=14336 keepinitrd mem=128M console=tty0 boot_mmc=y"
06:22.11phongredbootlinux
06:22.19phongredsorry spam
06:23.30phongredokie i see
06:23.59phongredadd "mem=128M"  in the file startup.txt
06:24.00phongredheheehehhe
06:24.01phongredthanks
06:25.09phongredu help me format SDcard
06:25.40phongredi can format with partionmagic
06:26.12phongred?
06:26.29phongredcan i Format with Partion Magic?
06:31.00phongredalo
07:00.12phongredalo
07:30.26Hymieyou again!
07:30.30HymieI've heard about you
07:30.36Hymietrouble, that's what I heard
07:31.00HymieI think you're out of luck until tomorrow (+5 hours at least from now)
07:34.09phongredi' here
07:34.21phongredalo
07:34.26phongredi'm here
07:34.36phongredare you there?
07:48.23*** part/#htc-linux phongred (n=phongred@203.113.144.83)
07:48.34*** join/#htc-linux phongred (n=phongred@203.113.144.83)
08:13.45*** join/#htc-linux RoEn_PC (n=roen@p54A65861.dip.t-dialin.net)
08:36.48*** join/#htc-linux goxboxlive (n=goxboxli@9.80-202-160.nextgentel.com)
09:06.34*** join/#htc-linux rmoravcik (n=rmoravci@icm6-orange.orange.sk)
09:11.13*** join/#htc-linux phongred (n=phongred@203.113.144.83)
09:11.21kimuSan^morning all
09:11.22phongredaloooooooooo
09:11.28phongredhello
09:11.33phongredlol
09:12.02phongredkimuSan^  <== u know use linux with Phone BlueAngel WM5
09:12.05phongred?
09:13.26phongredalo
09:13.44kimuSan^no
09:13.49kimuSan^trinity
09:14.10phongredlol
09:14.11phongredthanks
09:14.19phongredWho can help me?
09:20.10phongredalo
09:29.28*** join/#htc-linux pH5 (n=ph5@e178231216.adsl.alicedsl.de)
10:05.12*** join/#htc-linux VisN (n=VisualN@ip227.131.adsl.wplus.ru)
10:16.22*** join/#htc-linux webtom (n=chatzill@p54AB35FB.dip0.t-ipconnect.de)
10:26.43cr2kimuSan^: i've done some hermes tests in the morning. port0 is BT, port1 is GSM, port2 is IR
10:28.41kimuSan^ok
10:28.51kimuSan^have you put it on the wiki anywhere?
10:42.58*** join/#htc-linux Tjikkun (n=tjikkun@ip5451133d.direct-adsl.nl)
10:44.10*** join/#htc-linux jeanseb (n=jeanseb@88.164.32.155)
10:50.09*** join/#htc-linux LunohoD (n=alex@e180109105.adsl.alicedsl.de)
10:52.32*** join/#htc-linux rmoravci1 (n=rmoravci@icm7-orange.orange.sk)
11:18.35cr2kimuSan^: no, because i need to understand the clock frequency calculations.
11:19.51kimuSan^ok
11:21.32kimuSan^was it the uart clock it was for ?
11:22.47cr2the FCLK/n is selected for all 3 ports here, but i get the wrong frequencies with the divisors applied
11:23.16cr2you can do the following: switch BT, IR listen and phone on.
11:23.25cr2pd 0x50000000 0x30
11:23.29cr2pd 0x50004000 0x30
11:23.32cr2pd 0x50008000 0x30
11:23.45cr2and put the output into wiki.
11:24.19kimuSan^have you seen page 374 in the manual?
11:24.25kimuSan^there is a divisor description
11:25.19kimuSan^hmm my trin has some weird hangs sometimes when both wifi an bt is on.
11:27.16cr2page 378
11:28.36kimuSan^ok
11:30.00kimuSan^where on the wiki should I put it ?
11:31.44cr2hh.org ?
11:31.56*** join/#htc-linux rmoravcik (n=rmoravci@icm6-orange.orange.sk)
11:33.00kimuSan^yes but where in the masses of info :-)
11:40.09*** join/#htc-linux BabelO (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net)
11:41.48kimuSan^http://www.handhelds.org/moin/moin.cgi/HTCTrinity
11:43.36cr2ok.
11:43.51cr2we can refactor it later.
11:44.33cr2yes, it's the same config as on hermes.
11:46.33cr2what about s3c2410_platform_nand ?
11:46.49cr2<PROTECTED>
11:46.49cr2<PROTECTED>
11:46.49cr2<PROTECTED>
11:46.56cr2these fields.
11:50.28kimuSan^what about them ?
11:50.33cr2it is from NFCONF 0x4e000000
11:50.47cr2pd 0x4e000000
11:50.50cr2pd 0x4e000000 4
11:51.25kimuSan^4e000000 | 0000931c                            | ....
11:51.39cr2ok, trying to decode.
11:51.50cr2page 258
11:54.56cr21,3,1 vs. 1,6,2 on g500 ? hmm. how are the 15,30,30 calculated ?
11:55.15kimuSan^10 01 0 011 000 1 110 0
11:56.49kimuSan^what was HCLK
11:59.51kimuSan^hmmm HCLK should be 7.5?
12:00.19cr2drivers/mtd/nand/s3c2410.c
12:00.24kimuSan^but then tacle should be 2
12:00.30cr2FCLK,PCLK,HCLK ?
12:01.11cr2tacls=1, rph0=3, rph1=1
12:01.27cr2but how to they recalculate them to these numbers ?
12:01.39kimuSan^to the 15,30,30 ?
12:02.10cr2yes.
12:02.12cr2s3c_nand_calc_rate
12:03.16cr2#define NS_IN_KHZ 1000000
12:03.49cr2<PROTECTED>
12:03.50cr2<PROTECTED>
12:04.17cr2we look for "wanted".
12:04.25cr2we know the "result"
12:05.08cr2tacls = s3c_nand_calc_rate(plat->tacls, clkrate, tacls_max);
12:05.25cr2clkrate = clk_get_rate(info->clk);
12:06.25cr2even more complex than the pixclock calculations for pxafb ;)
12:06.55kimuSan^heh
12:07.02kimuSan^have you seen page 277?
12:07.02cr2s3c2410_nand_info *info
12:07.47cr2yes. can you dump the clock config too ?
12:08.05cr2CLKDIVN
12:09.14cr20x4c000014
12:09.45cr2the right decoding is actually a task for haret.
12:10.23cr2i've done it for the pxafb params. porting has never been more easy after that :)
12:10.29kimuSan^size is still 4?
12:10.49cr2just dump
12:11.00cr2pd 0x4c000000 0x20
12:11.03*** join/#htc-linux dothebart (n=willi@xdsl-84-44-229-87.netcologne.de)
12:11.18cr2size if always 4
12:11.29cr2unless you have some 16bit asic.
12:12.02kimuSan^4c000014 | 00000005 00000000 00000000 ffffffff | ................
12:13.39cr2page 294
12:14.32cr2oh man.
12:14.38cr2do
12:14.48cr2pd 0x4c000000 0x20
12:15.01cr2we need to know even more registers.
12:15.30cr2PCLK=HCLK/2
12:16.12cr2HCLK depends on CAMDIVN[9] FCLK/4 or FCLK/8
12:16.30cr2FCLK is 400MHz if i'm not mistaken.
12:16.40kimuSan^4c000000 | ffffffff 00033030 0003c042 0003fff0 | ....00..B.......
12:16.41kimuSan^4c000010 | 00000004 00000005 00000000 00000000 | ................
12:17.01kimuSan^yes
12:18.39cr2It is not easy to find a proper PLL value.
12:18.41cr2LOL
12:18.47cr2it's from the manual.
12:18.55kimuSan^hehe
12:19.35kimuSan^im off for a bit. need to get some food.
12:20.01cr2ok.
12:20.16cr2CAMDIVN=0
12:23.12kimuSan^so HCLK = FCLK/4
12:24.43cr2FCLK=400MHz, HCLK=100MHz, PCLK=50MHz
12:25.29*** join/#htc-linux g3gg0_ (n=g3gg0@ppp-82-135-0-250.dynamic.mnet-online.de)
12:25.34*** part/#htc-linux g3gg0_ (n=g3gg0@ppp-82-135-0-250.dynamic.mnet-online.de)
12:25.43kimuSan^yeps makes sence.
12:26.07cr2(wanted * clk) / NS_IN_KHZ +1 = result
12:26.30cr2wanted * 100 + 1 = result
12:27.09cr2wanted = (result -1)/100.
12:27.11cr2strange.
12:27.39kimuSan^why ?
12:27.53kimuSan^result was?
12:28.21cr2clk is in khz ?
12:28.32cr2then
12:28.35kimuSan^yes
12:28.50cr2wanted=(result-1)/100*1000
12:29.02cr2(result-1)*10
12:29.12cr2looks much better.
12:29.16kimuSan^yes
12:29.32cr21,3,1
12:29.41cr20,20,0 ?
12:30.22kimuSan^does that make sence ?
12:30.27cr2for .tacls, .rph0, .rph1
12:31.21cr2yes, but the g500 data seems to be wrong.
12:32.03kimuSan^oky
12:32.04cr2can happen, when things are not documented properly ;)
12:32.59kimuSan^:-)
12:33.51cr2ok.
12:34.15cr2hmm. it's for trinity.
12:34.25kimuSan^ok what?
12:34.30cr2i need to check the hermes values.
12:34.45cr2will use 0,20,0
12:35.42kimuSan^most likely it will be the same
12:44.38*** join/#htc-linux BabelOued (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net)
12:51.05cr2hermes realyy dislikes wrt54g
12:52.14cr2HaRET(1)# pd 0x4e000000 4
12:52.14cr24e000000 | 0000921c                            | ....
12:52.35cr2you have 931c
12:52.38kimuSan^yes
12:52.59kimuSan^so the value is not the same
12:53.01cr21,2,1
12:53.13cr2for me
12:53.16cr20,10,0
12:53.17kimuSan^0,10,0
12:53.20cr2then.
12:53.29cr2faster ? :)
12:53.34kimuSan^so not completely alike :-)
12:54.21cr2dumb hermes lost the connection again.
12:55.14kimuSan^nice
12:57.20kimuSan^hmm actually the trin likes my linksys better than any other APs I have tried.
12:57.36kimuSan^here it can be connected for hours and hours without disconnecting.
12:57.42cr2s3c24xx_init_clocks(0);
12:57.54cr2do we use the base 12MHz clock ?
12:58.05cr2still can't reconnect
12:59.05kimuSan^I had that issue once. needed to go into flight mode and back in order to get it working.
12:59.17kimuSan^like if it livelocked or somthing
12:59.46cr2i have bt on. too.
13:00.04cr2will switch it off.
13:00.23kimuSan^could be the problem. I have seen others complain about those two not being friends
13:00.42cr2they fight for hte radio.
13:02.57cr2maybe switch to the channel 14 ?
13:03.07cr2hermes may dislike it.
13:06.50maniacxsdoes the acx driver for universal work with wpa?
13:08.16kimuSan^mmm birthday cake.
13:09.38cr2maniacxs: was not working recently.
13:10.55maniacxsi tried to use wireless lan with wpa. i check the error messages ...
13:12.32cr2wep works.
13:12.45cr2kimuSan^: what is your default.txt ?
13:15.17maniacxscr2: i get "ioctl[SIOCSIWENCODEEXT]: Operation not supported" trying to use wpa
13:15.35maniacxsso i assume it is not implemented yet
13:15.58maniacxsonly wanted to know if it is so, or if i am doing something wrong
13:16.46cr2imho it is not implemented
13:16.59cr2fed up with wince.
13:17.24cr2no text editor, no save as in exploiter.
13:17.57kimuSan^cr2:
13:18.18kimuSan^set ramaddr 0x30000000
13:18.18kimuSan^set ramsize 0x40000000
13:18.18kimuSan^set kernel "ztri"
13:18.18kimuSan^set mtype 1225
13:18.18kimuSan^set cmdline "mem=64M"
13:18.18kimuSan^set kernelcrc 1
13:18.20kimuSan^bootlinux
13:18.40maniacxscr2: me, too. When mic is going to work in linux I am going to switch immediately
13:18.50kimuSan^as you told me to last night :-)
13:19.59kimuSan^cr2: spb pocket plus adds "save as" to  PIE
13:20.16kimuSan^but personally I use minimo as browser
13:20.42kimuSan^a bit ram eating but nice interface and quite fast.
13:21.04maniacxsminimo is really really slow here with the universal not usable
13:22.03cr2kimuSan^: one more proof that wince is shit.
13:22.16kimuSan^the latest release gave quite a speedup here
13:22.21kimuSan^but yes still a slow-loader
13:22.41kimuSan^cr2: couldn't agree more :-)
13:23.02cr2maniacxs: the mic is working, you only need to switch it properly, and a userfreindly solution is not yet there.
13:23.28maniacxscr2: i mean with using gomunicator
13:23.29kimuSan^still need to see some nice apps for the gpe though and some smaller widgets. they are way to space-filling right now.
13:25.17maniacxsis it known if the Xfbdev is using the console keymap, or a X keymap?
13:26.19maniacxsi was not able to find xev. I was searching half the day yesterday.
13:27.29cr2i'm not a gpe guy. i passionately hate gnome.
13:29.41kimuSan^and mostly just as restricted as M$ window
13:30.15kimuSan^but hey Tast cant be argued about :-) as long as we all hate wince :-D
13:30.28maniacxs:)
13:30.47cr2one does not quarrel about the tastes :)
13:31.51cr2kimuSan^: i like win95 look and feel. but i was a linux/x11 user before 95 :)
13:32.17cr2even before win3.1
13:32.58cr2well, not linux/x11 before win3.1, but x11 before win3.1
13:32.59kimuSan^had my 10 years anneversery  last summer
13:33.44cr2i found win3.1 very primitive compared to x11r3 :)
13:33.55kimuSan^I havent had windows as primary OS since back then. had a second pc with win98 though.
13:34.16cr2same here as you can guess.
13:34.25cr2my productivity in windows=0.
13:34.29kimuSan^3.1/3.11 was primitive indeed. but actually some of the best "ideas" from MS came to life back then
13:34.56cr2i have no clue how to use word and excel. computer illiterate, you know :)
13:34.57kimuSan^the rest is just idea-piracy
13:35.16cr2intellectual property is theft ;)
13:35.45kimuSan^I have to learn windows XP, office etc. all over again since my employer requires me to use win as primary OS.
13:36.29cr2i work in the university. only real losers use windows here :)
13:36.37kimuSan^but I do all my programming in emacs (soon Im back in Vim) in cygwin when I use windows.
13:36.48cr2boot params: RAMADDR=30000000 RAMSIZE=07000000 MTYPE=1213 CMDLINE='mem=64M'
13:36.57cr2way too much for the ramsize
13:36.58kimuSan^trin?
13:37.02cr2hermes
13:37.04kimuSan^ok
13:37.30cr2wince5 misreports ramsize.
13:37.44cr2i think haret should hardcode 64M for wince5.
13:38.08cr2!=64M are hardware hacks, and they know better how to adapt.
13:38.48cr2ok.
13:39.25kimuSan^my publisher is getting annoyed by me being 1 month behind schedule :-)
13:39.38cr2it's normal :)
13:39.42kimuSan^yes
13:39.56cr2for him to be annoyed, and for an author to be late.
13:41.56kimuSan^well I was actually on schedule, but there was a schedule mixup so I have to write the longest chapter in the same time as the shortest chapter. there like 3x more pages in the longest chapter.
13:42.10cr2i see.
13:42.32kimuSan^so they had to bare with me being "delayed"
13:42.43maniacxscr2: with switching on mic did you mean in linux or for gomunicator?
13:44.17cr2maniacxs: you need to send some i2c commands to route the mic to the phone. it's not connected to the phone directly like on blueangel/magician/others.
13:44.53maniacxshas this to be done once? or for every phonecall?
13:44.55cr2there should be an alsa mixer kernel code for that, and gomunicator/whatever should use it.
13:45.42cr2i'm not an expert on kernel audio, so it takes time.
13:46.20cr2but it is well known what to do, i.e. not an issue of some undocumented stuff.
13:50.13cr2Allocated 297 pages (tags=00281000/32ca6000 kernel=00282000/32ca7000 initrd=003A8000/32e99000 index=003A8000/32e99000)
13:50.17cr2Built kernel tags area
13:50.17cr2Built page index
13:50.18cr2Video buffer at phys=10800000
13:50.18cr2preload=700@003A1000/33048000 sj=003A1000 stack=003A6000/32e9a000 data=003A9000/33007000 exec=3304816c
13:50.18cr2Reading 1193316 bytes...
13:50.18cr2Read complete
13:50.24cr2then it's dead.
13:52.07kimuSan^dang
13:53.04cr2i don't see
13:53.07cr2Launching to physical address
13:56.05*** join/#htc-linux yacc (n=andreas@h1634.serverkompetenz.net)
13:57.27*** join/#htc-linux posokolvsky (n=psokolov@238.usernat.ip.net.ua)
14:11.57*** join/#htc-linux rob_w (n=bob@p85.212.177.9.tisdip.tiscali.de)
14:30.52*** join/#htc-linux pleemans (n=peter@d51A5E612.access.telenet.be)
14:36.44*** join/#htc-linux rmoravcik (n=rmoravci@pc-3s0zt5w2e4y0vzmhnrzq3a21zqajzfw.users.student.utc.sk)
15:10.44*** join/#htc-linux BobOfDoom (i=ato@210-9-141-44.netspeed.com.au)
15:45.24*** join/#htc-linux BabelO (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net)
16:11.47*** join/#htc-linux webtom (n=chatzill@p54AB35FB.dip0.t-ipconnect.de)
16:13.52kimuSan^cr2: did you figure out more?
16:28.45*** join/#htc-linux lpotter (n=ljp@c210-49-127-23.rochd1.qld.optusnet.com.au)
16:42.53*** join/#htc-linux VisN (n=VisualN@ip227.131.adsl.wplus.ru)
16:43.39*** join/#htc-linux TripleDES (n=zsirc@89.129.136.254)
17:34.44*** join/#htc-linux TripleDES (n=zsirc@89.129.136.254)
18:43.57lkclkimuSan: teach your employer about free software instead, by installing colinux, cygwin and the like.
18:44.25lkcland openoffice.
18:45.53kimuSan^well actually we have dusins of linux laptops and servers
18:46.14kimuSan^in my office alone + alot in other offices.
18:47.00kimuSan^they just want to have everyone to use windows primarily including outlook (yikes this is the worst shit ever)
19:08.00*** join/#htc-linux rmoravci1 (n=rmoravci@pc-nbhcnceqpo8wov6n0m468h4u6ax6am4.users.student.utc.sk)
19:09.59cr2lkcl: can you boot g500 with hte latest haret ?
19:15.23Kevin2I've pushed: http://www.handhelds.org/~koconnor/haret/haret-20070107.exe  in case any one is interested in the latest.
19:17.27cr2Kevin2: is it possible to clean the framebuffer RAM (in asm) just before jumping to the zimage ?
19:18.05cr2Kevin2: the version i have thinks that the ram is 96MB. maybe because wince says so ;)
19:18.26cr2should we force 64MB on all wince5 targets ?
19:23.04*** join/#htc-linux goxboxlive (n=goxboxli@9.80-202-160.nextgentel.com)
19:27.17kimuSan^cr2: or 64M unless the user specifies it to be anything else
19:33.20*** join/#htc-linux |Ralith| (n=ralith@soggy202.drizzle.com)
19:40.30Kevin2cr2: Enable logging via the earlyharetlog.txt trick, and see what it says during the memory detection step.
19:41.05Kevin2I think on wince5 we should be ignoring the dwStoreSize parameter and just use dwTotalPhys.
19:41.37Kevin2Zeroing framebuffer ram should be straight forward. But what's the reason for it?
19:41.52Kevin2I'll be back later.
19:45.30*** join/#htc-linux lilliput_ (n=Touriste@82-47-214-88.stb.ubr11.brad.blueyonder.co.uk)
19:53.29*** join/#htc-linux rmoravcik (n=rmoravci@pc-nbhcnceqpo8wov6n0m468h4u6ax6am4.users.student.utc.sk)
20:25.38cr2posokolovsky: should the asic3_keys driver support switches ? or should i write the asic3_switches driver ?
20:34.44drnogah, finaly found out what wrecked my linux system (took me weeks to figuere it out :() but Im back in trying to get linux running on my hw6915
20:36.55cr2good.
20:36.56drnocr2: I got as far as loading the kernel but never got the NFS running
20:37.07drnowell the NFS on my computer
20:37.20cr2have you installed opie on miniSD ?
20:37.47drnowell, my main priority was getting my system in a workable state
20:38.06cr2do you have an usb cardreader ?
20:38.17drnoyes
20:38.30cr2have you partitioned the miniSD card ?
20:39.03drnonop, its back to fat so I atleast can listen to mp3's
20:39.43cr2create 1 fat and 1 linux partition.
20:40.18cr2put your mp3s, haret & co. to fat. untar opie to the linux partition.
20:40.26cr2then boot.
20:41.05drnook, I have to find my miniSD to SDcard converter though
20:41.14cr2bbl
20:44.06*** join/#htc-linux asylumed (n=asylumed@196.211.116.2)
20:52.50drnowell, Im still updating my compile system, so its best to wait with building opie from cvs
21:10.20cr2get the precompiled image for universal.
21:10.31drnoit will work?
21:12.21drnoI guess so obviously :)
21:13.21goxboxlivedmo: You could also just try out the bootstrap-image. It is just a clean linux image without any GUI. To test Touchscreen you can just run ts_calibrate.
21:16.36drnoOpie-uni-2.6.16-hh3-R0.2.tar.bz2 is the one Im donwloading now
21:17.54drnobtw, goxboxlive I could build another system, or is the bootstrap to specified for GPE/Opie?
21:18.32posokolvskyHi!
21:18.51goxboxlivebootstrap is not specified for any GUI systems, it's just a clean linux system.
21:19.24goxboxlivehi posokolvsky> please answer to cr2 about this:  cr2> posokolovsky: should the asic3_keys driver support switches ? or should i write the asic3_switches driver ?
21:19.52posokolvskygoxboxlive, just yesterday I closed lkcl's one year old ticker in OE bugtracker which had zillion patches attached. Now you mail, not even a ticket. If you want it to be processed 1 year, you're on the right track ;-)
21:19.56drnook, thx goxboxlive
21:20.36drnoIll stick to the Opie build for a while though, that would make debuffing a bit easier if I have to reproduce or tell you about a bug
21:21.03oppizKevin2 you around>
21:21.05oppiz?
21:22.07posokolvskycr2, goxboxlive: it should support switches, I'd guess. at least, start from that assumption, and if patches are not to hairy, let's have single keys-related driver.
21:22.28goxboxlive<posokolvsky> it was not ment for you. Hrw told me that there where some bugs for the Universal. I sent it to you because there where some opie stuff there, and also just to keep you informed.
21:23.21posokolvskygoxboxlive, ok, but don't expect to bright response ;-)
21:23.36posokolvskygoxboxlive, I'm more interested in you VT-related fixes for OPIE.
21:23.41posokolvskygoxboxlive, are they in that patch?
21:27.30goxboxlive<posokolvsky> No it isent. As you see, it is in the htcuniversal specific folder, but i guess it's also affects other devices.
21:28.25goxboxlive<posokolvsky> DOnt you want me to send you a copy in the future?
21:30.11posokolvskygoxboxlive, yes, I *want* to be in cc:, but you could really ease our (adn your ;-) )life if they were single mail - single patch. And btw, there's *bug tracker*. But cc: anyway, I'm glad to see any improvments, and work towards getting them upstream.
21:33.17cr2posokolovsky: ok, i'll look at the switches. but it seems that i need one more switch that is available in linux/input.h Namely the NIGHT_SENSOR.
21:35.14goxboxliveposokolvsky> Well the bugtracker isent actually effective, look at the dates of my bug reports, and you see why. But this was not ment to be a patch or anything, it was just to inform hrw how i manage to build for htc universal.
21:35.44cr2drno: lkcl was using the universal opie image on sable. expect some minor bugs (serial ports) here and there.
21:36.49cr2but we wil iron them out RSN. and don't follow the BA way. creating more incompatible images is the last thing we really want.
21:37.11posokolvskygoxboxlive, you're bugs are well and safe, and will be processed. (you see, even one year old patches are processed ;-) ). Otherwise, I see, indeed, I heard hrw does great job of regression-testing builds for all machines.
21:37.25drno:)
21:37.55BabelOcr2: why don't follow the BA way ? :) :)
21:38.29cr2goxboxlive: i can't compile the opie for universal with yesterdays tree, because it wants some 18-hhXXX kernel instead of the current CVS.
21:39.35cr2BabelO: because rob_w created the BA config, it is completely outdated now and full of weirdness interesting only in historical aspect :)
21:40.30BabelOcr2: that s why i only use universal image, that is always up to date, thanks goxboxlive.
21:40.53BabelOcr2: AT+CBC command always return 0,0 on BA
21:41.31cr2BabelO: you have different serial port layout.
21:41.39drnocr2 are there any special options I chould pass to tar to extract it to the miniSD card?
21:41.57cr2which is not respected by the "universal" config. the same for sable and other devices.
21:42.20BabelOcr2: yes but all is done in init.d script :)
21:42.32cr2universal does not support AT+CBC at all, afaik. you have much better ways to get the battery status from /sys
21:43.06cr2BabelO: we need a case (${MACHINE}) in the init.d script.
21:43.45cr2drno: tar xfvjp file.tar.bz2 -C /where_is_your_linux_partition.
21:43.49BabelOcr2: yes i m agree with you, i use AT+CSQ for RF level
21:43.49goxboxliveBabelO: That was good to hear, you welcome, at OE it just get scolding for my work.
21:44.16drnok thx
21:44.19cr2BabelO: the HTCCSQ may be a better option, but you need to gather wince statistics
21:45.49BabelOcr2: yes i see, and if i monitor ttyS0 for a while, i see HTCCSQ sometimes with +CMT: ????
21:46.05cr2BabelO: i guess AT+CREG=2 works good ?
21:46.15BabelOyes :)
21:46.20cr2+CMT ?
21:47.00cr2then you can add a base station logger :)
21:47.06BabelOyes +CMT, i don't know what it is +CMT: 154,.... with some hexa code
21:47.20goxboxlivecr2: That's probaly because of an outdated defconfig file you have. It seems like you are using a old defconfig file (.18).
21:48.05goxboxlivecr2: i'll send u a mail
21:48.16cr2http://wiki.xda-developers.com/index.php?pagename=HTC_GSM
21:48.23cr2also observed on Universal.
21:49.35cr2goxboxlive: why can't it use the kernel defconfig. i'll spam the OE tracker with my defconfig updates ;-)
21:49.55cr2IMHVO it is silly.
21:50.39cr2for a tagged release use the OE defconfig, and for the CVS version the kernel defconfig.
21:51.00goxboxlivecr2: Read the mail i sent you.
21:51.15cr2not yet there.
21:52.50cr2goxboxlive: i'm also working on the 'other' leds driver, but it's a bit involved because of the current LED driver architecture.
21:53.25cr2and i'm thinking about a minimal update of the ALSA mixer code.
21:53.34cr2to support the builtin phone.
21:53.51goxboxlivegreat
21:54.18drnocr2 its booting with a few failures though, like not able to find the modules (obviously because I have a different kernel then the one the images asumes)
22:01.42cr2drno: yes. we need an automated solution for this problem. i'm inclining towards copying the modules from initrd, which is created at the same time as the kernel.
22:02.35cr2goxboxlive: i have a looked at your patches. some of them have zero chance to be ever applied. in some other cases it is not the best way to solve the problem.
22:04.55goxboxlivecr2: Yes i know that, i have also said that in mail to the OE devs. But they wouldnt let us build from CVS, so that we have to define in our local.conf. One of the reciver of the mail sent me an answer, i can forward it to you, with my comment to hin and the you see more of the history.
22:11.00cr2you want to patch angstrom-2007.1.conf
22:11.03cr2why ?
22:11.22cr2SRCDATE = "20071029"
22:11.34cr2PREFERRED_VERSION_linux-handhelds-2.6_htcuniversal = "2.6.19-hh7+cvs{SRCDATE]"
22:11.59cr2this is creating an endless patch sequence.
22:12.15cr2and they will never do it anyway.
22:12.46goxboxlivecr2: Read the mail, i dont want to pacth the distro conf file
22:12.56cr2let's create our own distro file. and will track the angstrom-2007.1.conf if appropriate.
22:13.28cr2but we need some way to build our image without 10000 patches.
22:13.42goxboxliveThis is how i build from cvs, i will remove it and add it to loacl.conf.
22:14.03goxboxliveWe are not going to have 10000000 patches
22:14.09cr2conf/machine/htcuniversal.conf
22:14.27cr2it's our own file, we can do whatever we like, it can be applied immediately.
22:14.42goxboxliveWhen we have most of the stuff working e can tag a version in angstrom.conf, but i dont see any poit of doing that now.
22:15.07goxboxliveThey dont want to let is build from cvs.
22:15.27goxboxliveIf you want to buld from cvs, you add it to your local.conf
22:15.28cr2gomunicator- i don't care.
22:15.33goxboxliveme neither
22:15.37cr2euro patch.
22:15.48goxboxliveI use gomunicator today just for fun, and testing
22:15.49cr2keep it as it is, because it is correct
22:16.14cr2the problem of € is the missing symbol in the console font. afaik.
22:16.32cr2can we load "our" console font ?
22:17.14cr2libopie2_cvs. ok it should be local.
22:17.17goxboxliveok, but it gives a failure message, so i cant include it as long as it doesnt work. When € works in the console i'll add it again.
22:17.23goxboxliveyes i agree
22:17.39cr2which failure message ?
22:17.51goxboxliveWhen loading keymap.map
22:17.59cr2strange.
22:18.03goxboxlivewhy?
22:18.18cr2i didn't see the euro, but did not have a load failure too.
22:18.34goxboxliveIf you define something that doesnt exists it gives you a failure message
22:18.44cr2because altgr-e is €
22:18.58cr2what does not exist ?
22:19.19cr2e ? :)
22:19.28cr2or altgr ?
22:20.26cr2packages/linux/linux-handhelds-2.6/htcuniversal/defconfig
22:20.58cr2i think it's wrong in principle. for a non-release build, which we are doing.
22:21.26cr2i wish you have deleted this file away.
22:22.13cr2packages/opie-taskbar/opie-taskbar/htcuniversal/opie
22:22.13goxboxlivecr2: IN case you add to kernelkeymap: keycode 2 = cr2 . It will give you a failure message when you load it. It's same with euro, it is missing somewhere, or maybe it's not "euro".
22:22.36cr2it's our own file, we are free to change it.
22:22.45cr2hmm.
22:23.01cr2what do you feed to 'loadkeys' ?
22:24.08cr2packages/qte/qte-common_2.3.10.inc
22:24.24goxboxlivecr2: non-release build. yes i agree, but i dont know how to let bitbake choose the htcuniversal_defconfig
22:24.51goxboxlivewhat do you feed to 'loadkeys' ? I dont know what you mean. Take a look at the patch and the you see
22:25.00cr2this is also wrong. if you really want to achieve that, create a patch, that will revert the applied patches :)
22:25.20cr2ok, i'll look at the file, not the patch.
22:25.43goxboxlivecr2: I know, but i have no clue how to make the patch and let it be aplied right. I have tried, but it failes.
22:26.19goxboxlivethe last one was ment for qte-common.inc
22:27.32cr2the last one was for loadkeys.
22:27.51cr2packages/tasks/task-opie.bb - this is a bit confusing for me, i have no idea.
22:28.57cr2so i'll say pick out the parts which don't need any discussion, and submit them separately.
22:29.16cr2we will think about the others.
22:29.56cr2btw, do you create the "national" keymaps ? btw, why "english" is the only chioce at opie startup ?
22:30.07goxboxlivecr2: Yes i will, but i'll wait for comments from hrw. He said he will take a look at it tomorrow, and let me know wich ones i need to work more with.
22:30.20cr2ok.
22:30.45goxboxlivecr2: Beacuse ihavent included any other languages.
22:31.08goxboxliveAnd that the other languages arent up2date
22:32.23cr2ok.
22:33.05cr2i actually agree with the comments of koen, and we need our distro file to resolve the base problem.
22:33.26cr2without dirty hacks, mind you :)
22:35.46goxboxliveYes, of course i understand that. But, i domt know how to do it, and if i ask for help i'll be yelled at so, dont you see my point ? It's easier and more silent to have local hacks. :-)
22:36.25cr2ok, i have the code for 4 leds now, will be extended later.
22:36.48cr2i see, but it's not an excuse for submitting local hacks :)
22:36.55cr2keep them local.
22:38.34cr2BabelO: can you gather stats of +CMT: strings ? i have not found any docs about it.
22:38.48drnocr2:  whats the key combination so I can use the / sign? ive triet almost every combination but the right one
22:40.00cr2goxboxlive: if everybody will submit his local hacks, the OE will turn into unmaintainable mess. sometimes i think it is already :)
22:40.02kimuSan^nite folks
22:40.27goxboxlive:-) yes i know, i got an answer form hrw now.
22:41.33cr2drno: you should create a keymap.
22:41.52cr2hmm. i'll look at sable.
22:43.02cr2drno: shift-p ?
22:43.04BabelOcr2: yes, i m saving all trace now :) and waiting for another CMT
22:43.20drnohaha Id wish that would work, that gives me a P
22:43.23cr2drno: what do you have for shift-p in your current keymap ?
22:43.36cr2'dumpkeys'
22:44.13cr2drno: the keymap is really weird, so you need your own custom map.
22:45.04cr2drno: hmm. i think it should be altgr-p
22:45.04drnoyeah I know
22:45.32cr2then euro will be altgr-e, as it is on other devices.
22:45.46cr2how is it in wince ?
22:45.56drnoall I have is 'dumpkmap'
22:46.43drnoI couldnt find it in wince zo I used the touchscreen for that if I needed it
22:46.54cr2in the angstrom image ?
22:47.25cr2weird. i'll vote for altgr-p. at least it has some logic behind.
22:47.37drnoand dumpkmap give me a 'dumpkmap: /dev/vc/0: No such file or directory'
22:48.16cr2wrong busybox setup.
22:48.20*** join/#htc-linux posokolvsky (n=psokolov@238.usernat.ip.net.ua)
22:48.31cr2devfs is dead for 1000 years already.
22:49.56drnoive got the image from http://gnulinux.biz/files/Universal/sd-mmc/2.6.16/
22:50.07cr2don't use it!
22:50.12drnogah
22:50.24cr2goxboxlive: can you remove this outdated stuff ?
22:50.46cr2Universal/tmp/Angstrom
22:50.54drnowhere can I get the image you are talking about then?
22:51.13cr2http://gnulinux.biz/files/Universal/tmp/Angstrom/
22:51.32drnook, including the kernel image?
22:51.54cr2no. you should use your own kernel. this one is for universal.
22:52.03drnook
22:53.02goxboxlivecr2: ok
22:55.20cr2goxboxlive: IMHO, the opie image "should not" contain the modules.
22:55.37cr2we need some better way to do it.
22:55.58goxboxlivecr2: When will you have finished the alsamixer and the switches?
22:56.08goxboxliveI need to answer the OE devs about a tag
22:56.19cr2and leds. i'm working ;)
22:56.20goxboxliveWe must build from a tag, and not from cvs
22:56.32cr2the kernel ?
22:56.36goxboxliveyes
22:56.48cr2don't think so.
22:56.56goxboxlivethey say we have to
22:57.02cr2then we should not build the kernel with OE at all.
22:57.07goxboxlivebecause of distro release
22:57.12cr2we don't have a usable tag.
22:57.54goxboxliveI have thought about it, i have also googled for other ways to build, but it seems so compicated
22:57.55cr2let's aim at 1 feb.
22:58.01goxboxliveok
22:58.20cr2i have many other things to do. and will be away on 21-23
22:58.36cr2and need to prepare a report, poster and a talk.
22:58.37goxboxliveOf course, i understand
22:59.12goxboxliveno rush, but i am so tired of the OE stuff.
22:59.20cr2build the kernel is complicated ?
22:59.31goxboxliveNo, images
22:59.35goxboxlivecomplete images
22:59.44goxboxlivelike the ones we have today
22:59.47cr2build the image without a kernel and modules.
23:00.12cr2it does not make much sense for a non-flashable image.
23:00.15goxboxliveThen wich machine should i add?
23:00.50cr2just don't build the kernel package. machine ? universal for the distro purposes.
23:01.08goxboxlivehmm
23:01.12cr2the init.d scripts should read /proc/cpu to get the ${MACHINE}
23:01.40cr2and always refer to the ${MACHINE}
23:02.10cr2not just irattach /dev/ttyS2
23:02.42goxboxliveBut, if we dont make any kernel-modules and zImage we will have a problem later
23:02.45cr2but switch(${MACHINE} ; case "HTC Universal":
23:02.58cr2no.
23:03.07goxboxliveBecause, now, we have to use the same gcc compiler as the image where buildt with, otherwise it wouldnt start
23:03.43cr2i compile the kernel with
23:04.00cr2PATH=~/stuff/tmp/cross/bin/:$PATH CROSS_COMPILE=arm-angstrom-linux-gnueabi- make
23:04.26cr2so it's ok.
23:05.28cr2we can create a bash script that will compile the kernel, haret, initrd, modules and copy modules to initrd.
23:05.57cr2then you put this stuff on the wince partition.
23:06.07cr2we can even create a cab file for it.
23:06.56cr2i'll never use OE to compile the kernel ;-)
23:07.02cr2it#s too much pain.
23:07.32cr2and is really worthless unless you will flash this image like on zaurus.
23:09.41drnocr2: the arnstrong image hangs on Freeing init memory: 104K
23:10.05drnoarnsrom*
23:10.08cr2drno: which compiler did you use for the kernel ?
23:11.07drnoI used gcc if thats what you mean
23:11.16cr2which version ?
23:11.47cr2i guess you can't compile linux kernel with msvc :)
23:12.05drno4.1.1
23:12.15cr2eabi enabled ?
23:12.16drnowell, it ran fine with the other image
23:12.44drno(scratches head) eabi?
23:13.01cr2# CONFIG_AEABI is not set
23:13.04cr2damn.
23:13.10cr2i'll update the defconfig.
23:13.46drnowhat is eabi anyway?
23:15.27cr2ApplicationBinaryInterface
23:17.12cr2Kernel Features->Use ARM EABI
23:17.55cr2enable it, and recompile the kernel.
23:17.59drnok
23:22.21goxboxlivecr2: can not get any further because he is using another gcc compiler than the one the image where created with. I have runned in to that problem several times, because i use my desktop computer for OE, and my laptop for the kernel.
23:22.51goxboxlives/can/drno can/
23:23.46goxboxlivewell it's getting late. good night .
23:26.37cr2drno: do you have a new kernel ?
23:31.21drnoeh, no was grabbing coffey
23:31.45drnoand I need to get the latest build from CVS anyway
23:32.12cr2ok.
23:32.16cr2good night.
23:32.47drnonn
23:32.52drnoand thx for the help :)
23:33.07*** join/#htc-linux rmoravci1 (n=rmoravci@pc-3s0zt5w2e4y0vzmhnrzq3a21zqajzfw.users.student.utc.sk)

Generated by irclog2html.pl by Jeff Waugh - find it at freshmeat.net! Modified by Tim Riker to work with blootbot logs, split per channel, etc.