IRC log for #htc-linux on 20090906

00:05.30Bushmang'nite
00:06.18dcordes_tmzt, in board-htcblackstone-panel.c ?
00:10.20Reactor16i can see Operator name :D
00:10.28Reactor16can i make call
00:27.18*** join/#htc-linux AstainKicksLTxda (n=AstainHe@unaffiliated/astainhellbring)
00:41.04*** join/#htc-linux ellisway (n=ellis@94.30.45.128)
00:53.44dcordes_see you guys tomorrow
01:49.19*** join/#htc-linux sd00 (n=smac@82-39-194-231.cable.ubr02.jarr.blueyonder.co.uk)
01:49.29*** join/#htc-linux rafyvitto (i=rafyvitt@24.54.253.34)
02:03.39*** join/#htc-linux stickboy (n=anonymou@128.153.210.13)
02:09.59*** join/#htc-linux druidu (n=druidu@93.122.209.148)
02:17.53druidunote on the errors xperia got when compiling the kernel (undefined reference to msm_proc_comm in board-saphire-panel.c)
02:18.12druiduit's caused by syncing with the latest code from upstream
02:18.30druiduthey added a new board, sapphire, which is by default enabled
02:18.59tmztright
02:19.07tmztdisable that for now
02:19.13*** join/#htc-linux unholy (n=IceChat7@24.54.253.34)
02:19.33druiduI only fixed htcraphael_defconfig... for other configs, when you load them, because CONFIG_MACH_SAPPHIRE is not present in the respective config files, it will be enabled by default
02:19.37tmztany idea on handover failing?
02:19.53unholyhey guys do any of you know if the deafault.txt command to overclock work on kaiser with msm7201?
02:20.04druidubut because proc_comm_wince excludes proc_comm (it's a choice item, only one possible)... it's impossible to support both sapphire and our boards
02:21.09tmztunholy: only if the kernel supports tat cmd line option
02:21.11druiduso maintainers of htcblackstone and htckovsky should update their configs using htcraphael_defconfig as a starting point (sorry I didn't mention it clearer in the commit)
02:21.25tmztharet just sends it to the kernel
02:21.56druiduand I say it again, htcraphael_android_defconfig is outdated so it should be either updated or removed (right now you can run android with htcraphael_defconfig)
02:22.01druiduhandover failing?
02:22.15druiduright when the console should switch from my console to framebuffer console?
02:22.34tmztyes
02:22.53tmztwhat hardwaew do you have?
02:22.58druidudiam100
02:23.04druiduit fails for me too sometimes
02:23.09tmztonly?
02:23.13druiduI think it's just like `bad luck` :)
02:23.16tmztyeah, it always has
02:23.33druiducouldn't find the cause... I am almost sure it's just the screen that's failing, the rest boots normally
02:23.33unholyhow do i know if the command supports it? its the nwest one if not mistaken for the vogue  that works on kaiser i guess
02:23.48druiduthat's the tat cmd line option?
02:24.18tmztI don't know what dzo's supports?
02:24.23unholykernel* no command lol
02:24.33unholynot*
02:24.52druidups: this is the commit I was talking about: http://git.linuxtogo.org/?p=groups/mobile-linux/kernel.git;a=commit;h=8b67377dc4d81aa0bb75a55dbd231c87fd9fc22c
02:25.11druiduah, lol
02:25.36tmztcr2 says it's that dzo changed board-raphael.h
02:25.51tmztdcordes split itto a daim.h
02:25.56tmztdiam.h
02:26.07tmztand diam.c
02:26.28tmzt<PROTECTED>
02:26.29tmzthttp://git.linuxtogo.org/?p=groups/mo
02:26.48druiduhm, cr2 sent me the patch to add board-htcdiamond.h
02:27.14druiduand also, I reverted the change that broke diamond (the stacked memory stuff)
02:28.29druidutmzt: when you get crashes, you should check the ram console buffer after restarting (via reset button)
02:28.56tmztI haven't tried it yet, I have cdma raph
02:29.02druiduit's @ 0xe0000 of size 0x20000
02:30.09druiduyou can dump it with haret like this: pwf log.txt 0xe000d 0x020000
02:30.16druidu(that address survives a reboot)
02:30.37druidu0xe000d because first 13 bytes are the circular buffer header
02:32.15druiduthe ram console is pretty cool 'cause it records even kernel panic message, everything up to the point the kernel freezes completely
02:32.36tmztyeah
02:32.45tmztwe should have used it
02:33.10druiduactually, I never thought there are sections of ram that survive a reboot :)
02:33.11tmztdiam boots now?
02:34.36druidumy diam boots, audio works (only out)... but I send the cid via the cmd line option I added (used dzo's initial value, autodetection fails for me)
02:34.56druidulockups at console switch are pretty rare
02:35.13druiduand android seems quite stable too
02:35.14tmztcid for adsp?
02:35.17druiduyes
02:35.24tmzt3d?
02:35.38druiduhaven't tried it actually, don't think it was enabled in config
02:36.11tmztyour not using dzo's config?
02:37.11druidugood idea, I'll try it now
02:37.12druidunope
02:37.21druiduusing htcraphael_defconfig
02:37.29druiduand my own cooked android :P
02:37.46druidudon't have a data connection though, haven't looked into it too much
02:39.55druidubtw, the ram console with early init option activated will probably not work, that memory area is not mapped
02:40.11tmztyeah
02:40.28druiduwith early init, you have it give it a virtual address, we pass a physical address via normal init and it works
02:44.36*** join/#htc-linux root (n=root@rgnb-5d87df12.pool.mediaWays.net)
02:45.32*** join/#htc-linux marcin88 (n=marcin@chello089078160132.chello.pl)
03:02.04druiduget this when compiling haret with mingw32ce toolchain: /tmp/ccLQAqVD.s: Assembler messages:
03:02.04druidu/tmp/ccLQAqVD.s:2371: Error: Rn must not overlap other operands -- `swpb r4,r0,[r4]'
03:02.04druidu/tmp/ccLQAqVD.s:2381: Error: Rn must not overlap other operands -- `swp r4,r0,[r4]'
03:04.02tmztno ides
04:16.43*** join/#htc-linux droid0011 (n=g1@p4FDCDDCC.dip.t-dialin.net)
04:59.10*** join/#htc-linux MrPippy (n=pip@75.11.160.45)
06:14.21cr2morning
06:51.20*** join/#htc-linux cr2 (n=cr2@ip-77-25-245-170.web.vodafone.de)
07:30.54dzomorning cr2, i just pushed a patch to add Audio Parameters and fix audio during calls.
07:31.28*** join/#htc-linux pH5 (n=ph5@g229162124.adsl.alicedsl.de)
07:31.29dzoSorry about the smem patch I thought raph had 64M too.
07:31.56dzoare you sure it's not there, perhaps its just protected.
07:37.05cr2dzo: raph has 288MB according to all specs, and writing above 32MB can't be read back
07:39.04dzoOK, i thought all 7201As had 64M, at least you have the extra 128M.
07:39.32cr2yes
07:39.39cr2does the mic work now ?
07:40.08dzoyes, it does for me.
07:40.40cr2ok, i'll try
07:41.36cr2BT floods the driver with irqs, never tried wifi yet and gps has some rpc problem...
07:43.58cr2-       info->max_task_id = 16;
07:44.00cr2+       info->max_task_id = QDSP_RTOS_MAX_TASK_ID;
07:44.17cr2i have also added this patch to adsp.c
07:44.42cr2it healed some problems, but created some others. (in dmesg)
07:45.31cr2playing .mp3 as mp3 does not work, but wince does not use it too.
07:46.09cr2sorry adsp.c means adsp_5200.c
07:46.26cr2but it's the same for other adsp_*.c
07:47.41dzook, the power collapse works a little but it's a bit of a hack and still needs some work.
07:47.41dzoi'm going to be really busy for the next few days but will try to get something done after that,
07:48.43cr2ok, thanks
07:49.37cr2you save some gpio conf in power collapse, i didn't understand the syntax, but isn't it diamond-specific ?
07:58.25*** join/#htc-linux kudIaty00 (n=barth666@klonowa1.jota.za.digi.pl)
07:58.33kudIaty00hi
08:10.24*** join/#htc-linux skodde (n=skodde@unaffiliated/skodde)
08:11.44*** join/#htc-linux Echo31 (n=olivier@mir31-4-82-240-194-54.fbx.proxad.net)
08:37.52Echo31cr2: Hi
08:38.10Echo31PH5: Hi
08:39.08pH5hej Echo31
08:44.04*** join/#htc-linux Zinbolic (n=zinbolic@84.238.80.215)
08:49.55tmzthey
08:51.46Echo31cr2: I  need some info about V05
08:54.28pH5hej tmzt
09:14.32*** join/#htc-linux x29a_ (n=x29a@unaffiliated/x29a)
09:14.42x29a_pH5: hi there, im trying to build my own rootFS for the magician, could you tell me which toolchain you were using (oe i guess?) and do you have some recent kernel configs?
09:29.07pH5hej x29a_
09:29.30x29a_howdy
09:29.34pH5yes, I'm using an oe toolchain (usually oe.dev angstrom)
09:30.39pH5I haven't built anything userspace related for a long time.
09:31.17pH5But my kernel playground (at git.linuxtogo.org) always contains a current defconfig
09:31.40x29a_ah sweet
09:32.46*** join/#htc-linux druidu (n=druidu@93.122.209.148)
09:32.54druidushit, there was a thunderstorm last night
09:33.14druiduit fried some of my isp's equipment, net and tv is completely gone
09:33.25pH5ouch
09:33.32x29ainsurance?
09:34.10druidunope, I wasn't affected my anything just that they don't work anymore :)
09:34.56druiduthey said 12 points went off-line at the same time last night
09:37.28druidulucky me I have internet via mobile
09:37.44druidubut now I'll disconnect each time I'll boot android :P
09:40.23x29aduh
09:40.28cr2lol
09:40.54cr2druidu: recording sound does not work for me. have your tried it ?
09:42.22*** join/#htc-linux luc_ (n=luc@89-115-128-35.cl.ipv4ilink.net)
09:50.24*** join/#htc-linux sd00 (n=smac@82-39-194-231.cable.ubr02.jarr.blueyonder.co.uk)
09:51.05cr2reboot does not work
09:54.50druidujust fetched the patch, gonna try now
09:54.55druiduh3d too
09:55.07druidureboot didn't ever work, poweroff does :)
09:55.15druidubtw, how can you reboot in android?
09:57.25druidualso, been trying to compile haret... it fails when compiling src/l1trace.cpp complaining about swp and swpb instructions: Error: Rn must not overlap other operands -- `swpb r4,r0,[r4]'
09:57.52druiduasm("swpb %0, %1, [%2]"  : "=r" (readval)  : "r" (writeval), "r" (newaddr));
09:57.55druiduany ideas? :)
09:58.32cr2which gcc is that ?
09:58.33druiduI guess it assigned r4 to both readval and newaddr vars, don't know how to tell it to use distinct regs for those
09:58.44cr2my older one does not have any problems
09:58.56druiduarm-mingw32ce-gcc
09:59.00druidulatest from cegcc
09:59.34cr2gcc version 4.1.0
09:59.36cr2here
10:00.05cr2Kevin2: ping
10:07.22druiduI found a solution
10:07.32druiduI'll just comment the freaking code, don't need to simulate an access fault
10:07.56cr2ok
10:08.27cr2we may actually separate haret from the bootloader
10:09.20druiduit's pretty easy to build a boot loader, just just have to load those files in memory at proper addresses, build some tags and go ahead
10:10.20cr2can be just a separate app inside the haret tree
10:10.29cr2with a better gui
10:11.02cr2druidu: what do you think about wifi ?
10:11.11druiduhmm, interesting, and with a cfg file to specify a list of options and values... so you can use the gui to create default.txt file
10:11.18cr2i think that it's the lowest hanging fruit now.
10:11.33druiduand the next useful thing to have :)
10:11.45cr2i'd prefer gps :)
10:12.15cr2and bt is also fun because of fm radio (for tmc)
10:13.18druidugps is like extra, it's not that crucial for the goal of dumping wince, I for one could live without it (but would prefer not to, of course)
10:13.26druiduI have an FM radio chip on my diamond
10:13.31druiduknow anything about that?
10:13.45druiduit's in the bluetooth chip?
10:13.49cr2we need bt
10:13.52cr2yes
10:14.05druiduso, we have power controls for bt (rfkill)
10:14.07cr2gps has some problems with rpc.
10:14.08druidudo we have them for wifi?
10:14.25cr2probably i've done something wrong in the servers/callbakc chain
10:14.49cr2the powerup code for wifi should be already there too
10:15.03cr2i think the problem was the firmware loader
10:15.27cr2and then there was a second wifi driver, so i was confused in the end
10:15.42cr2what to use, and what to do
10:16.05cr2maybe checking the wifi for kaiser is the best thing to do
10:16.06druiduhow does the wifi loader work? we have like a kernel driver and userspace loads it up via some /sys interface?
10:16.09*** join/#htc-linux Squarc (n=Squarc@82-217-32-29.cable.quicknet.nl)
10:16.22cr2no, it's firmware loader.
10:16.55cr2you need to load firmware to the wifi chip
10:17.31cr2the sdio is only for the eeprom and init/detect phase
10:18.11cr2there should be some mmaped io registers, but i'm not quite sure where they are
10:18.14druiduwhere's  the code for that in the android tree?
10:18.24druiduit's the same chip, we should have the base code for it
10:18.26cr2an older fw loader used mmap(/dev/mem)
10:18.52cr2the old g1 tiacx driver was in a separate git repo
10:19.04cr2"too big" to be included into the kernel tree
10:19.28druidu<PROTECTED>
10:19.36druiduso we need to get that calibration data from spl?
10:20.03druidu/dev/mem is not writeable anymore? or is that /dev/kmem?
10:20.48cr2calibration data is in nand
10:22.58*** join/#htc-linux leaigor (n=laigor@188.134.36.14)
10:23.23druiducan we read nand?
10:23.29cr2yes
10:23.45cr2don't enable it with android though :)
10:24.03druiduI see bootloader needs to to a lot of work
10:24.17druidua lot of tags to send
10:24.20cr2why ?
10:24.30druiduATAG_MSM_PARTITION
10:24.35druiduATAG_SMI
10:24.40cr2it's just obfuscated google way of doing things
10:24.41druiduATAG_MSM_WIFI
10:24.58druidutrue, so we could do all that in the kernel I suppose
10:25.18cr2or even in userspace
10:25.20druiduI guess source code to amss was too big of a corporate secret to give out as open source
10:25.30cr2hehe
10:26.18cr2http://opencellid.org
10:27.07druidunone in romania :)
10:27.34druiduaren't they like I don't know... thousands of them in a country?
10:28.48druidubrb, gonna test the audio in patch, and h3d
10:29.06druidu(net via mobile)
10:31.53*** join/#htc-linux MrPippy (n=pip@adsl-75-11-160-45.dsl.sndg02.sbcglobal.net)
10:35.47cr2hi MrPippy
10:36.39*** join/#htc-linux cnossa1 (n=cnossa@a213-22-39-95.cpe.netcabo.pt)
10:36.39tmztwell, if we started with something like jumpspl and used it to jump to a bootloader (no longer in ce)
10:36.55tmztwe could use that for flashing as well at some point
10:37.38cr2tmzt: the first stage to do is booting via resume
10:38.02cr2once it will be working, you can talk about flashing
10:38.17cr2tmzt: where is your 6150 patch ? i've lost the link
10:38.48MrPippyhey cr2
10:39.06cr2MrPippy: tell us more about wifi :)
10:39.09tmztpeople.openezx.org/tmzt
10:39.15tmztsomewhere
10:41.35cr2tmzt: the rpc version is always =0 on wince amss
10:42.16MrPippythe wlan driver seems to load correctly when i use a patch from the kaiser guys (addisu), but i don't think they had good performance with it
10:42.41MrPippyand i don't have an android userland to the point where it can do anything with the wifi, it just resets when i check the box to try enable it
10:43.06cr2MrPippy: which driver are you using ? where one can look at the source ?
10:43.22cr2you should not need android to run wifi
10:43.38MrPippydriver is in android's git, system/wlan/ti/sta_dk_4_0_4_32
10:44.06cr2ok
10:44.44MrPippyand with this patch everything loads ok but it's really slow (like 5 minutes just to upload the firmware) http://forum.xda-developers.com/showthread.php?p=4261210#post4261210
10:47.49MrPippyafter you load the module, you have to run wlan_loader with a firmware file (Fw1251r1c.bin, get it from \windows), a tiwlan.ini file (i got mine from a kaiser image,  theres also one in android's git), and a calibration file that contains the MAC address and other info and is stored in the eeprom
10:48.50cr2i think eeprom is stored in nand
10:48.51MrPippyi enabled the MSM MTD device and dumped the whole thing out to find the calibration info, there's a pretty unique header at the start but it occurs twice in the eeprom, and the second one worked
10:48.59cr2we can just rip it out of there
10:49.10cr2ok
10:49.14MrPippyand after wlan_loader runs successfully, tiwlan0 shows up as an interface
10:49.40MrPippyyeah i'm not sure if the offset is gonna be the same for all devices or not
10:50.25MrPippyright now i was trying to sync up to the latest kernel version, the early boot messages show up but then it's just a white screen, and it resets maybe 20 seconds after that
10:51.01cr2http://www.htc-linux.org/wiki/index.php?title=RaphaelNAND
10:51.10tmztwe should really work on the new driver
10:51.25tmztwl12xx
10:51.25cr2this is not the latest kernel (with WSOD)
10:51.44*** join/#htc-linux dcordes (n=luke@f054180100.adsl.alicedsl.de)
10:52.14*** join/#htc-linux druidu (n=druidu@93.122.135.18)
10:52.18MrPippydoes android support wl12xx?
10:52.18*** join/#htc-linux dzo (n=dzo@121.98.128.127)
10:52.43druiduso it's not working... calls don't work at all (rild crashes when call is incoming, or when dialing)
10:52.47tmztno
10:52.59druidustrange though, it used to work
10:53.18druiduas for h3d, don't think we mapped the proper memory for it to work
10:53.53druiduD/SurfaceFlinger(   97): pid 97 requesting gpu core (owner = -1)
10:53.53druiduW/SurfaceFlinger(   97): couldn't grant gpu core to pid 97
10:53.53druiduD/EGL     (   97): requestGPU returned -1
10:54.03tmztbut that's libhardware and wpa-supplicant
10:54.08druidubtw, how did they sign haret? so it doesn't get that warning when run?
10:54.31cr2disable cert check
10:54.37cr2in the registry
10:54.47cr2or add your own cert
10:54.51cr2and sign with it.
10:55.27tmztsome things need to be in a specific gid to work
10:55.32tmztin android
10:55.42druiduhow do I sign the .exe?
10:57.49cr2druidu: it was documented in some itsutils docs
10:57.50*** part/#htc-linux sd00 (n=smac@82-39-194-231.cable.ubr02.jarr.blueyonder.co.uk)
10:58.32dcordeshi
10:59.05dcordesdruidu, do you think the freeze error on kovsky and blackstone might be ram console related?
11:00.57druidufound signtool.exe, trying with that
11:01.05druidudcordes: hey, what freeze error?
11:01.25cr2druidu: i'd like to know how to do this with free software
11:01.49dcordes<PROTECTED>
11:01.49dcordes<PROTECTED>
11:01.49dcordes<PROTECTED>
11:01.49dcordes<PROTECTED>
11:01.49dcordes<PROTECTED>
11:01.49dcordes<PROTECTED>
11:01.51dcordes<PROTECTED>
11:01.53dcordes<PROTECTED>
11:01.55dcordes<PROTECTED>
11:01.57dcordes<PROTECTED>
11:01.59dcordes<PROTECTED>
11:02.01dcordes<PROTECTED>
11:02.05dcordes<PROTECTED>
11:02.07dcordes<PROTECTED>
11:02.09dcordes<PROTECTED>
11:02.11dcordes<PROTECTED>
11:02.13dcordes<PROTECTED>
11:02.15dcordes<PROTECTED>
11:02.17dcordesthis occurs on kovsky as well as on blackstone
11:02.24druiduhmm
11:02.28druidulemme check my log
11:02.44druidu[    2.527496] io scheduler anticipatory registered (default)
11:02.44druidu[    2.528167] mddi: init() base=0xc6010000 irq=16
11:02.49druiduit's mddi related I guess
11:02.54cr2dcordes: disable ram console. that's easy to check
11:03.02druiduit can't be ram console
11:03.05dcordesok
11:03.06*** join/#htc-linux AstainHellbring (n=AstainHe@unaffiliated/astainhellbring)
11:03.12druiduit would crash either before that either at all
11:03.20MrPippyok, on diam500 this will extract the working calibration file: dd if=mtd0.img of=calibration bs=1 count=752 skip=33161280
11:03.25druiduit's definitely the mddi & panel init code that's crashing
11:03.55druiduMrPippy: how did you get mtd0.img ?
11:03.59dcordesdruidu, could it be that board-htcblackstone-panel.c didn't keep up with other changes ?
11:04.18druidu[    2.528167] mddi: init() base=0xc6010000 irq=16 is in mddi.c
11:04.23druidudoesn't get to panel I think...
11:04.30cr233161280-0x2400000
11:04.35druiduthe mddi code calls the other stuff
11:04.50dcordesdruidu, so where must I look at?
11:05.12MrPippyi built in the MSM MTD support and used dd to copy it from /dev/mtd/mtd0ro, its a 256 MB file
11:05.14druidusee mddi_probe() in drivers/video/msm/mddi.c
11:05.20druidu<PROTECTED>
11:05.22druidu<PROTECTED>
11:05.28druidu<PROTECTED>
11:05.35druidumust be one of these 3
11:06.04druidutry printk at start of mddi_probe and after each of these instructions
11:06.48druiducr2: just dump it from haret, is nand mapped to memory?
11:07.06cr2druidu: no.
11:07.32cr2druidu: there is no nand dump support in haret (so far)
11:08.00MrPippyextracting it straight from /dev/mtd/mtd0ro should work, thats what kaiser guys do in their init file
11:09.21druiduso why did I brick my device when I tried to read from nand? :)
11:09.42druidudon't remember exactly, maybe i tried to mount it and it didn't recognize something and started to write stuff in there...
11:09.44cr2druidu: android wrote there
11:10.05cr2and since the spl is at the beginning, it was damaged
11:10.12druidudon't see why it would write
11:10.22druiduthe mount didn't succeede
11:10.37cr2which mount ?
11:10.49druiduwhen I tried to mount mtd
11:11.04druiduit never made it, was generating a lot of debug output
11:11.08cr2you can't , there is not "real" filesystem ther
11:11.09druiduand at some point I reset
11:11.10cr2e
11:11.24druiduoh well, I'll try again :)
11:12.11dcordesdruidu, ok trying that now
11:14.10cr2druidu: you need to provide the mtd params on the cmdline
11:14.27cr2the nand size is 512MB - 36
11:14.34cr2hmm. diam has 256MB ?
11:16.17*** join/#htc-linux sd00 (n=smac@82-39-194-231.cable.ubr02.jarr.blueyonder.co.uk)
11:17.05MrPippyyeah i got a 256MB file when i copied it, but that might just be diam500
11:19.09druidupdadb.net :)
11:19.29*** join/#htc-linux br1ck (n=br1ck@xdslch131.osnanet.de)
11:20.14druiduROM+type:  Flash EEPROM    ROM-capacity:  256 MiB
11:21.51MrPippyi've been running with mtype=2040 (htcdiamond_cdma) so that usb gets initialized right, but i also have to change htcraphael-mmc.c to use the gsm mmc pdata, since diam500 has internal storage like diam100, not a microsd slot like diam400
11:22.19MrPippythere's got to be some cleaner way to do that--probably don't want to go back to a msmsdcc_id style kernel cmd line option, but i don't know if theres any way to autodetect it in the kernel
11:23.15dcordescr2, startgps works?
11:24.40dcordesdruidu, it seems to hang exactly after the 3 instructions. the last thing I see are my 4 printks.
11:25.13dcordesdruidu, http://pastebin.ca/1556002 mddi: dcordes: platform_get_resource2 is the last line
11:27.40druidumddi: init() base=0x%p irq=%d
11:27.44druiduthis is the first you don't see?
11:28.51druidu        mddi->irq = resource->start;
11:28.58druiduthere's no way this would crash
11:30.23dcordesyes, this one is not visibile. mddi: dcordes: platform_get_resource2 is the last line printed.
11:31.21dcordesmaybe something about irq changed?
11:33.16druiduthere's no call to irq functions in there...
11:33.25druiduget resource just accesses a structure
11:34.31dcordesdruidu, so? I have no clue what to make of this
11:34.51dcordesdruidu, are there other init processes running simultaneously?
11:37.40druidulemme check some stuff
11:38.23druiduproblem is... it doesn't make sense
11:38.37druiduonly explanation is that resource points to junk, but it's not null
11:38.57druiduyou see log from ram console or on the screen?
11:39.40dcordeson screen
11:39.43druiduactually you don't even have ram console enabled
11:39.44druidutry that
11:39.55dcordesok
11:40.12MrPippyi just booted the latest kernel, didn't get the WSOD but it did panic, i think on the battery driver
11:40.39druiduhttp://pastebin.ca/1556012
11:40.41druiduthis is the code
11:40.54druiduadd it to board-htcblackstone.c
11:41.06druiduthe structures +         &ram_console_device, in the device list
11:41.21dcordesok
11:41.21*** join/#htc-linux marex (n=marex@thor.hackndev.com)
11:42.17dcordesdruidu, oh board-htckovsky.c I use. btw, the config I use is exact arch/arm/configs/htckovsky_defconfig only with the sapphire removed
11:43.33dcordesdruidu, where in the code must the section in the paste be added?
11:43.42druidusame :)
11:43.53druidujust above the devices array
11:43.55dcordesbefore initdata ?
11:44.08druidustatic struct platform_device *devices[] __initdata = {
11:44.17druiduput the ram_console first in the device list
11:44.19druiduso it's loaded first
11:46.30dcordesdruidu, http://pastebin.ca/1556017
11:47.39Bushmanhttp://www.youtube.com/watch?v=1nhOGmWC6ZE
11:47.43Bushmanmy new pet
11:47.46Bushmanhavin' a dinner
11:48.04dcordesBushman, it eats htc phones running linux?
11:48.21dcordesdruidu, ok it compiles and boots with the patch
11:48.50Bushmandcordes: no... it eats bugs
11:49.14dcordesdruidu, now softreset and read ramconsole using haret?
11:49.15Bushmangets quite handy when you'r doingsome coding stuff :P
11:50.29druiduyes
11:50.55druidupwf log.txt 0xe000d 0x20000
11:51.05druidu(first 13 bytes are the log header, binary junk)
11:51.07*** join/#htc-linux Zoolooc (n=fredsiba@p549553CA.dip.t-dialin.net)
11:54.56dcordesok done
11:56.38druiduso, what's the crash?
11:56.43dcordesdruidu, how to look at the log?
11:56.55dcordesI think it's binary only
11:56.57dcordesno text
11:57.04druiduthen the ramconsole didn't work...
11:57.08druiduit should be plain text
11:57.18druiduopen it anyway, try vi or smth
11:57.26druiduor run `strings` on it
11:57.28dcordesdruidu, ok. did you look at my patch above?
11:57.45druidupatch is ok
11:57.54dcordesit has nothing useful..
11:57.55dcordes$ strings log.txt
11:57.55dcordesH6E#
11:57.55dcordes$%&'
11:57.59druiduram console enabled in config?
11:58.03dcordesno
11:58.05dcordes:)
11:58.08druiduwithout the early boot flag, with the verbose flag
11:58.15druiduit's under devices -> staging -> android
11:58.37druidus/devices/drivers/
12:01.24dcordesstaging ?
12:01.59druidusorry
12:02.02druiduthat's in 2.6.29
12:02.09druidudevice drivers -> misc
12:02.41druiduso make sure ram console is enabled, with verbose msgs, no error correction, no start ram console early
12:02.53dcordesok
12:02.54druidu(early won't work 'cause that memory is not mapped)
12:03.11druiduwere they enabled?
12:03.16dcordesno
12:03.21dcordesnothing
12:04.41druidualso
12:04.43druiduadd a printk
12:04.51druidubefore this
12:05.20druiduprintk(KERN_INFO "irq_resource=%p\n", resource);
12:05.41dcordesok
12:08.26dcordesdruidu, ok the new log looks exactly the same
12:09.15dcordesdruidu, where is  printk(KERN_INFO "irq_resource=%p\n", resource); ?
12:10.23druidusorry, I meant before this:         mddi->irq = resource->start;
12:10.27druiduadd that line
12:10.30dcordes$ strings log_02.txt
12:10.30dcordesH6E#
12:10.30dcordes$%&'
12:10.37druiduit seems that resources->start causes the hang
12:10.53druiduif ram console wasn't enabled in config it will surely not work
12:11.17druidulog.txt should definitely be a text file, exactly all you see in the console
12:11.29dcordeslog_02.txt was dumped after a boot with ram console enable.d
12:12.22dcordesnow I have double printk
12:12.24dcordesprintk(KERN_INFO "mddi: dcordes: platform_get_resource2\n");
12:12.24dcordesprintk(KERN_INFO "irq_resource=%p\n", resource);
12:12.24dcordesmddi->irq = resource->start;
12:12.36druiduok, let's see what that does
12:12.55druiduram console is initialized before mddi, I think, should have worked...
12:12.56dcordesCONFIG_ANDROID_RAM_CONSOLE=y
12:12.56dcordesCONFIG_ANDROID_RAM_CONSOLE_ENABLE_VERBOSE=y
12:12.56dcordes# CONFIG_ANDROID_RAM_CONSOLE_ERROR_CORRECTION is not set
12:12.56dcordes# CONFIG_ANDROID_RAM_CONSOLE_EARLY_INIT is not set
12:13.10dcordesbut it didn't .
12:13.48druidudo you get any new message on the screen?
12:13.53druidu<PROTECTED>
12:13.53druidu<PROTECTED>
12:14.01druiduthis should be printed when ram console loads
12:14.17dcordestbh I cannot process the stuff that was
12:14.18dcordesfast
12:14.23druidu<PROTECTED>
12:14.24druiduthis too
12:14.29dcordesI will look for it
12:14.38dcordesshould be pretty early no?
12:14.57druidutry kernel hacking -> Delay each boot printk message by N milliseconds
12:15.18druidunope, sorry
12:15.21druiduI see it's after mddi
12:15.46druiduwait
12:21.34dcordesdruidu, okok
12:21.41druidugonna test smth, brb
12:21.45druidusee what that printk prints
12:36.06*** join/#htc-linux marcin88 (n=marcin@chello089078160132.chello.pl)
12:43.35*** join/#htc-linux druidu (n=druidu@93.122.135.1)
12:43.46druiduok, found a way to enable ram console much earlier
12:43.57druidudrivers/misc/ram_console.c, at end of file
12:44.07druidu//module_init(ram_console_module_init);
12:44.07druidufs_initcall(ram_console_module_init);
12:44.18druiduchange module_init into fs_initcall
12:44.30druiduthis will initialize ram_console before any other device driver
12:45.43druiduso after reboot we'll see the cause of the panic
12:52.46*** join/#htc-linux eway (n=ellis@94.30.45.128)
12:53.27dcordeseway, hi there
12:53.55*** join/#htc-linux kimhoon (n=kimhoon@s559116c1.adsl.wanadoo.nl)
12:55.19*** join/#htc-linux druidu (n=druidu@93.122.135.18)
12:55.35dcordesdruidu, I added it. If I have seen right there was ram_console visible
12:55.43dcordesdruidu, and the last line is irq_resource=asdjasöd
12:55.57druiducan't find the app that disabled certificate signing checking
12:56.08druidunothing after that?
12:56.11dcordesnope
12:56.12druiduin ram console
12:56.16dcordesnononoo
12:56.17*** join/#htc-linux MrPippy (n=pip@adsl-75-11-160-45.dsl.sndg02.sbcglobal.net)
12:56.17dcordesdisplay.
12:56.21druidutry ram console now
12:56.23druidushould be there
12:56.24dcordesyes
13:00.07dcordes$ strings log_03.txt
13:00.08dcordesH6E#
13:00.08dcordes$%&'
13:00.37dcordespwf log.txt 0xe000d 0x20000 is the correct command?
13:01.22dcordes0xe000d is the correct base?
13:01.49dcordes#define MSM_LOG_BASE 0x0e0000
13:01.50dcordes#define MSM_LOG_SIZE 0x020000
13:05.18*** join/#htc-linux rosseaux (i=znc@84.19.167.129)
13:07.31dcordesdruidu, no worky
13:13.03druiduyes
13:13.06druiduthat's the one
13:13.21druidud is to skip first 13 bytes
13:14.07druidudunno, maybe e0000 doesn't survive reboot in your diamond
13:14.17druidudzo told me that one
13:15.44druiduwhat's the address printed after irq_resource= ?
13:16.56dcordesdruidu, I don't have diamond but kovsky. I will reboot to check for address
13:17.33dcordesram_console got buffer at e0000
13:17.40dcordes...
13:17.41dcordes..
13:17.54dcordesirq_resource=c028a500
13:18.57dcordesaha
13:19.23dcordesram_console: no valid data in buffer (sig = 0x00000000
13:19.44dcordesconsole [ram-1] enabled
13:21.05druidunope, it's definitely not working, the log should out in the clear
13:21.51druiduthat's not an error, it's ok (it saves the old log there, if it finds it)
13:22.36dcordesok
13:22.39druiduyou rebooted using a soft reboot right?
13:22.42druiduno poweroff
13:22.49dcordesoh wait
13:23.01dcordesnot 100%
13:23.09druiduwhat do you mean?
13:23.18dcordesI'm not 100% sure. I softresetted now
13:23.23druidudoes kovsky have like the red reset button?
13:23.26dcordesand will try once again
13:23.49dcordesno kovsky has a black reset button that lives under the backcover.
13:23.52druiduunder the stylus
13:23.58druiduah
13:24.12druiduas long as it's soft reset, data should be left in ram intact
13:25.54dcordesit is
13:26.23dcordeshow can we assure that on kovsky/blackstone the ram we use is not overwritten?
13:26.32dcordes..on wince boot.
13:28.20AstainHellbringits ram of course its over written reset should dump everything in ram
13:28.42ali12341it doesnt though
13:29.01druidutry to zero out the memory
13:29.07druiduit's protected on my diamond
13:29.07ali12341finding a bit that doesn't get trashed by WM is tricky though
13:29.39dcordesis the address overwritten on raphael?
13:30.05druidunope, don't think so
13:30.33druidunow... normally nothing should happen so zeroing out that memory should fuck up anything
13:30.38druidui just get a protection error
13:30.48dcordesdiamond, raphael vs kovsky, blackstone have different memory setup
13:30.51druiduwhich is probably why is not overwritten
13:30.54dcordesafaik
13:30.58druiduI see
13:31.00dcordeskovs blac have more I think
13:31.04druiduwell, gotta find a protected one
13:31.17druiduask dzo how he found this one
13:31.23dcordesdzo, any hints?
13:32.18druiduI hate thunderstorms...
13:33.37*** join/#htc-linux ellisway (n=ellis@94.30.45.128)
13:34.43*** join/#htc-linux druidu (n=druidu@93.122.135.18)
13:34.57dcordesdruidu, no lightning rod present?
13:35.13dcordesor does it screw up your internets?
13:35.22*** join/#htc-linux druidu (n=druidu@93.122.135.18)
13:35.24*** join/#htc-linux ImCoKeMaN (n=imcokema@pool-74-99-149-138.hrbgpa.fios.verizon.net)
13:42.02dcordes~seen dwaradzyn
13:42.03aptdwaradzyn <n=chatzill@chello089079067084.chello.pl> was last seen on IRC in channel #htc-linux, 16h 15m 33s ago, saying: 'it looks like that: http://tinypic.com/r/6h6e4i/3'.
13:47.31*** join/#htc-linux eway (n=ellis@94.30.45.128)
13:49.01*** join/#htc-linux toi (n=toi@d54C2AAB7.access.telenet.be)
13:50.19*** join/#htc-linux luc_ (n=luc@89-115-128-35.cl.ipv4ilink.net)
13:52.03dcordesbblö
13:59.36*** join/#htc-linux druidu (n=druidu@93.122.135.1)
14:00.19*** join/#htc-linux druidu (n=druidu@93.122.135.1)
14:00.36druiduwhere can I find that damn signature verification disabling app? :(
14:00.48druidubeen looking for that for hours and no luck... I know I had it some time
14:04.14*** join/#htc-linux Reactor16 (n=Reactor1@41.105.117.128)
14:05.17*** join/#htc-linux druidu (n=druidu@93.122.224.233)
14:05.29Reactor16hi all
14:07.45druidumy kingdom for the cert/sign verif disabling tool...
14:10.56*** join/#htc-linux ellisway (n=ellis@94.30.45.128)
14:15.48ali12341druidu: for WM?
14:17.13Reactor16any one tried to play with Ril ?
14:19.34ali12341druidu: this? http://xman892.blogspot.com/2007/02/how-to-series-windows-mobile-security.html
14:21.01*** join/#htc-linux eway (n=ellis@94.30.45.128)
14:28.28Reactor16how do i to remap keyboard for Htc VOX ?
14:28.37Reactor16i use wing linux
14:29.36druiduyeah, WM
14:29.48druidumanaged to sign the file actually, gonna see if it worked
14:29.59druiduosslsigncode - open source alternative cr2 :)
14:30.27*** join/#htc-linux ellisway (n=ellis@94.30.45.128)
14:35.52*** join/#htc-linux druidu (n=druidu@93.122.195.66)
14:36.17druiduarghh... I'm going mad, I signed the fucking file but of course, my certificate is self-signed, it's not trusted
14:36.28druiduthere was a cab that disabled the security check
14:36.39druiduso you could just run unsigned .exe files, or install unsigned cabs
14:37.27druiduali12341: not that, it was a wince .cab that I would just install on the device
14:37.37ali12341sdkcerts.cab?
14:42.33druiduactually the message itself is: the file cannot be opened. either it is not signed with a trusted certificate, or one of its components cannot be found
14:42.45druiducould it be that my compiled harets needs some dll?
14:43.19druidusdkcerts.cab... don't know, what does it do?
14:45.13druiducan someone test my compiled haret.exe, maybe it's something with my device... ?
14:46.50mickeylmorning
14:50.13druiduI hate windows mobile, from the bottom of my heart...
14:51.33mickeyl<PROTECTED>
14:51.43mickeylHEAD seems broken here
14:54.36mickeylyou better #ifdef CONFIG_ADSP the call to snd_set_device in htc_hw.c
14:54.39*** join/#htc-linux druidu (n=druidu@93.122.135.1)
14:54.56mickeylor have said hardware autoselect it
14:55.16druiduargghh... what does one have to do to get a fucking haret compiled and running???
15:00.06Reactor16any tried to run on wing linux sim-locked phone ?
15:03.14*** join/#htc-linux pigeon__ (n=pigeon@eth5284.nsw.adsl.internode.on.net)
15:08.35*** join/#htc-linux Alex[sp3dev] (n=alex_dfr@ip-95-221-34-14.bb.netbynet.ru)
15:09.31*** join/#htc-linux FR^2 (n=frquadra@frquadrat.de)
15:11.34*** join/#htc-linux xperia (n=chatzill@77-56-91-148.dclient.hispeed.ch)
15:11.53xperiahello
15:13.04xperiajust wanted to tell that the htc kovsky build is broken. i cant boot it. for me it looks like that the changes in the panel are the reason for the problem
15:13.18*** join/#htc-linux ellisway (n=ellis@94.30.45.128)
15:13.56*** join/#htc-linux miknix (n=miknix@gentoo/developer/miknix)
15:14.56xperiadruidu: do you have problems compiling haret ?
15:16.40xperiaP.S. i have the same freeze problems on the htc-kovsky device like decordes !
15:17.51Alex[sp3dev]ok, do you mean dzo's commit to linuxtogo? the one you suppose to have broken xperia?
15:19.29druiduI managed to compile it
15:19.33druidujust can't get it to run on my device
15:19.52druidueither because it's unsigned, or because something's missing (a dll ? shouldn't be any missing...)
15:20.10xperiadruidu: i had something similar. i new flash of the rom has removed the problems with not certified software !
15:20.51xperiai think this is windows that blocks haret to run.
15:22.02druiduyeah, struggling to make a ca, get a ca-signed certificate and sign the mother fucker...
15:22.15druiduopenssl is the tool I hate the most :)
15:23.08xperiaalex:yeah exactly. the new patches for the lcd panel screen.
15:23.24Alex[sp3dev]druidu, why not sign with microsoft test certificate (from sdk)?
15:23.49Alex[sp3dev]xperia: ok, i'll try to apply the patch and see what is what)
15:24.14xperiaalex: can you compile and boot the htc-kovsky build. i assume you cant
15:25.10Alex[sp3dev]well.. does only panel break compilation or sth else? just don't want to download all the patches
15:25.53xperiathe compilation isnt breaked. just it freeze if you try to boot it.
15:26.16Alex[sp3dev]ok. then i'll try it
15:26.33*** join/#htc-linux dcordes-kais (n=dcordes-@ip-77-24-104-115.web.vodafone.de)
15:26.40dcordes-kaisxperia hi
15:26.42Alex[sp3dev]some raphael-specific stuff was removed from panel as far as i see.
15:27.29dcordes-kaisxperia can you please read the channel logs? links are in topic
15:27.44mickeylmorning dcordes-kais
15:28.32dcordes-kaismickeyl moin
15:29.04dcordes-kaismickeyl good times are coming :)
15:29.14mickeylhmm
15:30.40mickeyli remain sceptic, but crossing fingers.
15:30.58*** join/#htc-linux druidu (n=druidu@93.122.135.1)
15:31.56*** join/#htc-linux LpBv (n=quassel@93.2.190.117)
15:32.06LpBvHi
15:32.08*** join/#htc-linux druidu (n=druidu@93.122.135.1)
15:32.16dcordes-kaisLpBv hi
15:32.31druiduAlex[sp3dev]: I'd do just about anything right now, just tell me how :D (i'm a n00b when it comes to this stuff)
15:32.41LpBvIt's the first time, I m connecting here
15:33.02dcordes-kaisLpBv welcome then
15:33.22druiduit may be that it's not that it is unsigned, but some dll is missing... :-/
15:33.32druiduanybody volunteers to test my haret.exe?
15:33.43LpBvI ve a problem with the last package posted by TaC on a touch  pro Raph100
15:33.45Alex[sp3dev]druidu: ok, i'll send you the needed files. it is a windows tool, though, so i'll check if it works under wine. anyway, show us your haret)
15:33.53*** join/#htc-linux Gnutoo (n=gnutoo@ABordeaux-152-1-84-218.w86-196.abo.wanadoo.fr)
15:34.17dcordes-kaisdruidu  just exec it?
15:34.52dcordes-kaisLpBv what sort of package?
15:35.02druiduthe message I get is: "the file haret.exe cannot be opened. either it is not signed with a trusted certificate, or one of its components cannot be found"
15:35.26LpBvthe android package he posted this morning
15:35.26druiduso i don't even know which of the two is the problem
15:35.31LpBvon xda
15:35.43dcordes-kaisdruidu got it available somewhere? I can try on kovsky w/ tmob nl stock rom
15:36.03Gnutoohi if I put code for pin I've CME error 30
15:36.14Gnutoothat is an error about the network
15:36.15LpBvhttp://forum.xda-developers.com/showthread.php?t=402002&page=363
15:38.21xperiadcordes-kais: hello decordes. do you have resolved the problem with the kovsky freeze. cant find in the logs any info for the problem.
15:38.33xperiahow to solve it
15:39.10*** join/#htc-linux kudIaty00 (n=barth666@klonowa1.jota.za.digi.pl)
15:40.21druidubrb
15:40.33Reactor16any one tried to run wing linux on sim-locked phone ?
15:41.55xperiaReactor16: the most of the user here are developing and testing on htc-raph htc-blackstone htc-kovsky
15:42.58xperiawing linux is something that i have heared only a few times in the last months.
15:43.41dcordes-kaisxperia everything I tried with druidu's help is in the logs. don't know what you want
15:43.42Alex[sp3dev]it should run without any problems, if you use haret (most likely they use it on wing, no one boots via spl debug image these days)
15:45.54*** join/#htc-linux druidu (n=druidu@93.122.135.1)
15:45.56*** join/#htc-linux kw (n=nnscript@adsl-178-60-36.gnv.bellsouth.net)
15:46.07Reactor16what is distro u work with ?
15:46.24kwwhen i boot hen new TAC i just get a gray screen or it reboots, it all varies by the mtype i have a diam500 what should the mtype be?
15:46.35xperiadcordes-kais: i have reaeded the logs. there where some dicussion about a ram console that didnt worked on the htc-kovsky right becouse probably of memory overwritings. my build freeze likes your if i am not wrong. can you boot with the new kernel angstrom on htc kovsky ?
15:48.04dcordes-kaisxperia jeez I even posted my boot error log in the channel after putting thr pastebin link like 3 times...
15:48.43dcordes-kaisxperia jeez I even posted my boot error log in the channel after putting thr pastebin link like 3 times...
15:49.07dcordes-kaiskw ask tac?
15:49.45kweh.
15:49.47kw:(
15:49.51kwi try 1805
15:49.59kwand it gives me a gray screen when it goes to load the image
15:50.42xperiadcordes-kais: i am like a computer. understand only 1==yes or 0==no every other things give me a syntax error :-) from you answer i assume the problem is not solved.
15:51.00dcordes-kaiskw look nobody knows who tac is and which kernel source/revision he uses for his packs
15:51.19dcordes-kaisxperia 2
15:52.31xperiakw: the right mtype is in a file in the sources. the link was posted yesterday here in the chanell by druidu. its somehing like mach/asm/mtypes.c try a search like git grep mtype
15:52.49kwNot running linux =\
15:53.25xperiathat sound like you are a second class inhabitant :-)
15:53.48kwWhat's that supposed to mean exactly?
15:54.22xperiafirst class inhabitant == linux users second class inhabitants==windows users
15:54.45xperiaps. the files can be found with gitweb on linuxtogo
15:54.52kwaye
15:56.16*** join/#htc-linux druidu (n=druidu@93.122.135.1)
15:56.41dcordes-kaiskw I propose using a known working pack until your builder fixes the new one
15:57.08kwThe only pack i know that works is extremely old.
15:57.39dcordes-kais:(
15:57.44dcordes-kaisbye byye
16:04.32xperiakw>when i boot hen new TAC i just get a gray screen or it reboots, it all varies by the mtype i have a diam500 what should the mtype be?
16:04.33xperiahttp://git.linuxtogo.org/?p=groups/mobile-linux/kernel.git;a=blob;f=arch/arm/tools/mach-types;h=861bbb579bdaac465d0603c6db3cf90fc0fb5ecd;hb=refs/heads/htc-msm-2.6.27
16:06.08*** join/#htc-linux Zinbolic (n=zinbolic@84.238.80.215)
16:06.37*** join/#htc-linux MrPippy (n=pip@adsl-75-11-160-45.dsl.sndg02.sbcglobal.net)
16:06.37xperiakw: did you find now the mtype ?
16:07.39kw2040 ?
16:08.50xperiakw: thats for the cdma version. your mtype 1805 should be the right one
16:08.54kwAnd then it says "waiting for sdcard"
16:09.16kwI have a cdma sprint diamond
16:09.22*** join/#htc-linux druidu (n=druidu@93.122.135.1)
16:09.58druiduok, so I managed to get a working haret thanks to Alex[sp3dev]
16:10.11xperiakw: then 2040 is the right one for you. it look like that you have the same problems like other people here. a lot of builds are broken and dont boot !
16:10.25kwah
16:10.26kw:(
16:10.31*** part/#htc-linux XorA|gone (n=nnXorA@www.xora.org.uk)
16:10.38druiduI patched linboot.cpp to load initrd at kernel addr +50MB  (instead of +5 MB) and now it boots
16:11.38xperiadruidu: should i try your build for me kovsky kernel ?
16:11.56druiduneah, it's just in a really early stage
16:12.08druiduit crashes right after extracting initrd, gotta incrementally enable more drivers
16:14.25xperiaalex: did you have success ?
16:15.24Alex[sp3dev]xperia:no, was busy with haret and looking through commit logs. downloading the snapshot right now (well, i don't have time to clone the git right now)
16:16.36xperiaalex: ahh okay.
16:19.07*** join/#htc-linux cr2 (n=cr2@ip-77-25-246-243.web.vodafone.de)
16:25.12cr2druidu: the max_task_id=16 in adsp is really a bug. i have compared to wince, it uses 19 too.
16:25.21cr2where do i post the ptch ?
16:25.37druidupastebin?
16:25.50druiduI'm got expecting git apply to work, but what the hell :)
16:26.09cr2diff --git a/arch/arm/mach-msm/qdsp5/adsp_5200.c b/arch/arm/mach-msm/qdsp5/adsp_5200.c
16:26.10cr2index b1c8db5..d98c621 100644
16:26.12cr2--- a/arch/arm/mach-msm/qdsp5/adsp_5200.c
16:26.13cr2+++ b/arch/arm/mach-msm/qdsp5/adsp_5200.c
16:26.15cr2@@ -270,7 +270,7 @@ int adsp_init_info(struct adsp_info *info)
16:26.22cr2-       info->max_task_id = 16;
16:26.23cr2+       info->max_task_id = QDSP_RTOS_MAX_TASK_ID;
16:26.34cr2it's a one-line pat cch
16:27.41druidulol, ok :)
16:27.44druidudescription?
16:27.46cr2the same bug is in all androed kernels, somebody may submit it to their bugtracker
16:28.20druiduwell, is it serious?
16:28.32druiduandroid-based phones seem not to mind
16:28.46cr2i got an error messages in dmesg
16:29.57cr2[Sa Sep 5 2009] [21:02:06] <cr2>        [ 1858.555145] adsp: module AUDRECTASK: READY
16:29.58cr2[Sa Sep 5 2009] [21:02:06] <cr2>        [ 1858.564880] [RPC] REPLY to xid=41 @ 0:00000001
16:30.00cr2[Sa Sep 5 2009] [21:02:06] <cr2>        [ 1858.576232] audpre: CFG ENABLED
16:30.01cr2[Sa Sep 5 2009] [21:02:06] <cr2>        [ 1858.589172] adsp: bogus task id 17
16:30.03cr2[Sa Sep 5 2009] [21:02:06] <cr2>        [ 1858.597839] adsp: bogus task id 17
16:30.04cr2[Sa Sep 5 2009] [21:02:06] <cr2>        [ 1858.606292] adsp: bogus task id 17
16:30.06cr2[Sa Sep 5 2009] [21:02:06] <cr2>        [ 1858.614715] audrec: PARAM CFG DONE
16:30.07cr2[
16:30.24cr2it needed task 17, which is completely legitimate
16:30.38cr2AUDRECPREPROC
16:30.43cr2or something like that
16:31.07cr2but the max_id=16 says it's bogus
16:31.51cr2so i've looked at the wince implementation, asnd they compare against 0x13
16:32.13cr2=19, aka QDSP_RTOS_MAX_TASK_ID
16:32.37cr2if androids did not hit the bug, does not mean that it does not exist ;)
16:34.58cr2there are some other hardcoded constants in adsp_5200.c, but only googel people know what's their purpose
16:37.21cr2druidu: please also copy the adsp_5200.c to adsp_6150.c
16:37.56cr2we will eventually need to fix the module index and offsets for those poor CDMA people
16:38.44cr2then the #define WINCE will be = AMSS5200 || AMSS6150
16:51.31druiduadding MSM_AMSS_VERSION_WINCE_CDMA to Kconfig
16:54.19*** join/#htc-linux Echo31 (n=olivier@mir31-4-82-240-194-54.fbx.proxad.net)
16:54.41cr2druidu: 5200 already exists
16:54.58Echo31hi cr2
16:55.04cr2hi Echo31
16:58.29druiducr2: what do you mean?
16:59.10cr2default 5200 if MSM_AMSS_VERSION_WINCE
16:59.22cr2default 6210 if MSM_AMSS_VERSION_6210
16:59.33cr2imho, the most logical way is to do
16:59.47cr2default 5200 if MSM_AMSS_VERSION_5200
16:59.57cr2default 6150 if MSM_AMSS_VERSION_6150
17:00.29cr2the actual version can be picked by parsing the smem info data (at runtime)
17:01.02cr2MSM_AMSS_VERSION_WINCE == MSM_AMSS_VERSION_5200 || MSM_AMSS_VERSION_6150
17:01.14cr2something like that
17:01.26cr2the board_init function
17:01.56cr2should print the board id (fc048 byte for raph/diam, something else for raph2/diam2)
17:02.04cr2and the AMSS version
17:02.34cr2if the amss version does not match, a BIG warning is to be done :)
17:08.31*** join/#htc-linux balans (n=Administ@212-123-149-239.ip.telfort.nl)
17:09.30druiduok, gonna do a major find -replace in that folder :)
17:13.30cr2:)
17:19.38cr2245760 is hardcoded for PLL0. good.
17:22.42*** join/#htc-linux MethoS- (n=clemens@host-091-096-215-048.ewe-ip-backbone.de)
17:28.45*** join/#htc-linux nrirclog (n=nrirclog@netripper.com)
17:36.53*** join/#htc-linux osilvan (n=o@161.122.222.87.dynamic.jazztel.es)
17:43.58druidu2 new commits
17:43.58druiduhttp://git.linuxtogo.org/?p=groups/mobile-linux/kernel.git;a=shortlog;h=refs/heads/htc-msm-2.6.27
17:44.36druiduupdates raphael defconfig and done the amss 6150 stuff
17:44.47druidus/updates/updated/
17:45.38druidubetter read the descriptions and update your config files accordingly
17:46.36*** join/#htc-linux nrirclog (n=nrirclog@netripper.com)
17:52.18cr2druidu: thanks. looks very good, now i need to update the adsp numbers there
17:53.03cr2i think the RPC progs are the same
18:00.27*** join/#htc-linux toi (n=toi@d54C2AAB7.access.telenet.be)
18:00.53druidubtw, why don't you request commit access? :)
18:02.39cr2you do it better :)
18:02.50cr2and i can have time for other things
18:03.38cr2i think we can get rid of the "frequency" member in clock struct
18:04.36cr2it can be derived from the PLL frequency source (S) and the MNP numbers.
18:04.53cr2the PLL frequencies are hardcoded in acpuclock.c
18:05.15cr2if we were to change them, it will create some trouble for the ARM11 speed :)
18:06.04druidubtw, somebody posted smth on the forum thread about a cmd line param that overclocked arm11?
18:06.14druiduhave you seen that?
18:06.35cr2change the divisors, and you can overclock it.
18:06.45cr2i think underclocking is much more fun :)
18:07.01cr2like running at 19.2MHz clock speed 8)
18:07.16cr2because it saves a lot of power
18:07.31druiduyeah, maybe just when it's idle, but not when you're using it
18:07.53druidugreat, with the toolchain alex gave me, compile works like a charm, no errors, no nothing
18:08.19druiduthis is definitely not my day... and my lazy isp didn't fix the problem yet
18:08.38cr2druidu: what is PCB XC setting ?
18:08.54cr2the LCD power gpio depends on it.
18:10.44druidudon't know what that is
18:11.40cr2+0xfc048  4b  model id (XC)
18:11.48cr2i think it's 1b
18:11.52cr2wiki is wrong
18:12.26cr2and here you have the AMSS version:
18:12.30*** part/#htc-linux sd00 (n=smac@82-39-194-231.cable.ubr02.jarr.blueyonder.co.uk)
18:12.35cr2+0xfc030   g23 ver
18:12.45cr2a string pointer
18:12.47druidushared ram base + that?
18:12.52cr2yes
18:13.14cr2where is 0x01f00000 remapped ?
18:13.49cr2the query of model id (XC) belongs to the lcd init
18:14.44druidu#define MSM_SHARED_RAM_BASE   IOMEM(0xE0100000)
18:14.50cr2ok
18:15.00druidumodel id, on what board is that?
18:15.09druiduthe main board with msm chip?
18:16.01cr2raph/diam
18:16.06cr2+0xfc08c   softreset =0x55aa00ff
18:16.19cr2this may be important for reboot...
18:16.30druidulike... just write something there
18:16.32druidu?
18:16.52cr2like write this value, and send a DEX message
18:20.22druiduok, gonna try to write the AMSS version and PCB XC setting
18:20.29druidushould i do it in board_init for both?
18:20.36druiduactually
18:20.37cr2yes.
18:20.43druidubetter in proc_comm_wince.c
18:20.46druiduso it's in all boards
18:20.55druiduit's where I put all the debug crap
18:20.58cr2the PCB XC will be used on raph/diam LCD powerup
18:21.03cr2and down
18:21.11cr2check PCB XC
18:21.13cr2if (XC) gpio_0x63=1
18:21.14cr2else    gpio_0x52=1
18:21.38cr2do we detect the panel id now ?
18:22.24druiduok, be back in 15 gonna try it in haret then on the device
18:22.34druidupanel id, not sure
18:23.14druidubrb
18:26.09*** join/#htc-linux swc|666 (n=infidel2@unaffiliated/swc666/x-4934821)
18:36.05*** join/#htc-linux mallochigh (n=Administ@63.135.231.146)
18:36.30mallochighI have a kaiser, I am wondering about what provides sound for it? Is it the msm7200 chip? Or is it an aux chip?
18:37.45cr2mallochigh: most probably msm7200 (+dsp). but without the docs it's difficult to say
18:48.49*** join/#htc-linux druidu (n=druidu@93.122.135.1)
18:48.54druiduit's done
18:49.34mallochighwhats done?
18:50.17druiduthe debug stuff, PCB XC and AMSS version printing
18:50.21druidujust pushed the patch
18:50.46mallochigh{just a spectator}
18:50.48cr2druidu: it seems that they've cut'n'pasted the support for all possible toshiba outputs from codeaurora
18:51.09cr2we don't really need such an extensive mddi setup :)
18:51.29druiduhere's my dmesg: http://pastebin.com/m7e60b83e
18:51.38druiduPCB XC: 00000001
18:51.43druiduAMSS version: 52.29.25.12H
18:51.50*** join/#htc-linux astrolite (n=astrolit@c-d7c9e455.023-352-6c6b701.cust.bredbandsbolaget.se)
18:51.55mallochighJust a stupidity/discontinuity check, anyone working on msm7200 sound driver? Kaiser android folks really need it, right now it works but things are broken in it which prevent me from using sipdroid.
18:52.30cr2druidu: 52.* means 5200
18:52.47cr2druidu: the PCB XC is 1 byte
18:52.54cr21 or 0 actually
18:52.55druiduyeah, just copied 16 bytes and printed them as a string
18:53.13druiduXC, what does that stand for?
18:53.21cr2druidu: you can compare the amss version with the kernel compiled value
18:53.26cr2no idea
18:53.45cr2i suggest pcb_model_id
18:53.57druiduI dumped it from haret, that's it
18:54.43cr2ok. next question. what is your panel_id ?
18:54.46druiduRUU_Diamond_HTC_WWE_EastEurope_1.93.479.3_Radio_Signed_Diamond_52.29.25.12_1.00.25.05_Ship.exe
18:55.16druiduI keep forgetting that, it's right there, there's another field for it
18:55.20cr2---get_panel_type---
18:55.22cr2msleep 10
18:55.23cr2read GPIODATA(150000)
18:55.24cr2id=0
18:55.26cr2if bit 0x10 is not set -> id+=1
18:55.28cr2if bit 0x04 is not set -> id+=2
18:55.29cr2=0 -> panel id 5 (hitachi)
18:55.30cr2=1 -> panel id 0xa (sharp)
18:55.32cr2=2 -> panel id 0x11 (toppoly2)
18:55.34cr2=3 -> panel id 2 (toppoly)
18:55.35cr2are we doing it already ?
18:55.42druidudon
18:55.44druidu't think so
18:55.54druidumddi can identify that I think, in read caps
18:55.56druidugotta test smth
18:56.05cr2ok. should be the part of panel-init
18:56.13druidulemme play a bit with 2.6.29 and I'll post more patches later
18:56.16cr2read caps ?
18:56.30cr2it's for the mddi client chip, not the lcd
18:56.35cr2ok
18:57.05cr2i need to patch the adsp_6150 too
18:57.15druiduah, you're right
18:59.06druidubtw, can't we reset arm9?
18:59.44druiduif we could reset arm9 then we could be much closer to boot state
19:00.00druiducold boot state, i mean
19:00.53cr2we don't know yet
19:03.21*** join/#htc-linux iseeka (i=5aa86195@gateway/web/freenode/x-hnjgmboydqqbifbm)
19:09.20cr2druidu: this is a shadowed wince copy
19:09.23cr2+0x081034   panel_id
19:09.56cr2but you can't rely on it for standalone boot
19:11.25cr2yeah, the spi init tablez
19:12.30druiduah, thought so
19:12.39druiduin git, how do I revert only one file to HEAD?
19:12.50druidugit reset --hard works only for all files
19:13.06cr2=!git_expert
19:13.15cr2pH5 is not here
19:15.26druidugonna do it the lame way, add, commit the rest of the files, reset
19:18.06ali12341git checkout <file>
19:18.44druiduhmm, lol, good to know :P
19:26.27*** join/#htc-linux druidu (n=druidu@93.122.248.29)
19:30.51*** join/#htc-linux stickboy (n=anonymou@128.153.210.13)
19:34.34cr2druidu: 16bit values in the spi table ;)
19:35.20*** join/#htc-linux luc_ (n=luc@89-115-128-35.cl.ipv4ilink.net)
19:35.33cr2we need SPI_WRITE16 function
19:35.55*** part/#htc-linux mallochigh (n=Administ@63.135.231.146)
19:36.36cr2should check how to do it
19:36.37druiduwe write to SPI via microp right?
19:36.42cr2no
19:37.10cr2via mddi to toshiba and toshiba sends to lcd
19:37.15druiduah, ok
19:37.18cr2complicated.
19:37.35druidunot that much, just that we don't have the full specs, gotta reverse engineer them
19:37.57cr2you have hitachi lcd ?
19:38.33cr2yes, i'm just tired for today with lcd
19:39.14druiduI don't know, I keep forgetting that :P
19:39.20druiduI know it's different from raphael
19:39.40cr2hmm
19:39.41druiduand I can't run haret unless I activate activesync & disable net
19:39.52druiduraphael has sharp?
19:40.21cr2need to check my smem dump
19:40.39druiduaa, I have a smem dump saved
19:40.39druiduwait
19:42.19cr2full smem
19:43.21druidu+0x081034   panel_id
19:43.23druiduthis?
19:43.25druidudamn, I don't
19:43.43druiduit's 0x081034  physical, not in smem
19:43.57cr2i have 0x5
19:44.13druiduwhat address?
19:44.29cr2panel id 5 (hitachi)
19:44.47cr2yes 0x81034 physical
19:52.33*** join/#htc-linux druidu (n=druidu@93.122.135.1)
19:52.44druiduok, I dumped some memory too
19:52.49cr2:)
19:54.26druidu0xa mine
19:54.45cr2panel id 0xa (sharp)
19:54.59cr2different table
19:55.05druiduyep
19:55.31cr2you may try to implement the panel id detection code
19:59.21cr2druidu: write the 0x55aa00ff to +0xfc08c as i've said before setting the gpio0x19=0
19:59.35druidufor reset?
19:59.41cr2for reboot. maybe it will help :)
19:59.43druidubtw, have you seen the other projects @ https://www.codeaurora.org/gitweb/quic/la/
19:59.51cr2yes
20:00.04cr2i want the tvout driver code from them :)
20:00.06druiduthey also implemented some of the proprietary stuff from g1
20:01.03cr2druidu: but that's real hard shutdown above.
20:01.57cr2the soft shutdown sends DEX 0x14
20:02.48*** join/#htc-linux surge (i=surge@pool-98-118-154-21.bflony.fios.verizon.net)
20:03.56druidubut ram stays the same?
20:04.25cr2which ram ?
20:04.35druiduram memory
20:04.42druiduwhere does it jump to? spl?
20:04.50cr2sdram (EBI) is wiped
20:05.17druiduand again, how is the 0xe0000 area still there after soft reboot via red button?
20:06.38cr2it's different ram (SMI)
20:06.51cr2maybe it's powered in a different way
20:07.35cr2ok, for shtudown you send dex 0x14
20:08.14cr2if it times out, then you put this magic number, and pulls the gpio 0x19 down :)
20:08.25druidushutdown works
20:08.31cr2what are we doing right now ?
20:08.35cr2not for me
20:08.53druidurpc call
20:09.01cr2??
20:09.11druiduworks perfectly on my diam100, shutdown i mean
20:09.13druidupower off
20:09.37cr2=halt
20:09.44cr2i mean arm9 reboot
20:09.58druiduaaah, arm9 reboot
20:10.03*** join/#htc-linux dcordes-kais (n=dcordes-@ip-77-24-250-84.web.vodafone.de)
20:10.05druidudex 0x14 is for that?
20:10.14druidudamn, linuxtogo is down again
20:11.22druidu<PROTECTED>
20:11.32druiduin htcraphael_reset
20:12.00druiduaaaa, I see, so you wanna put  0x55aa00ff to +0xfc08c, now I get it
20:12.34cr2them
20:12.39cr2then msleep()
20:12.48cr2and pull the gpio
20:13.02druiduput should we put that before msm_proc_comm_wince() ?
20:13.10*** join/#htc-linux MethoS-- (n=clemens@host-091-096-210-148.ewe-ip-backbone.de)
20:13.13cr2after msleep()
20:13.30cr2dex+msleep+0x55aa00ff+gpio0x19
20:14.13druiduok, I'll try
20:14.41druiduso this will just restart arm9 right? so all rpc system is reinitialized
20:17.08druiduwhat about this? https://www.codeaurora.org/gitweb/quic/la/?p=kernel/msm.git;a=blob;f=arch/arm/mach-msm/reset_modem.c;h=5d5b0f9a06201a77331e4e719f5fce6a5f43af13;hb=refs/heads/android-msm-2.6.29b
20:18.04cr2different amss
20:18.45*** join/#htc-linux MethoS_ (n=clemens@host-091-097-241-141.ewe-ip-backbone.de)
20:20.27druidudo you think we should dump the google android tree and adpot the codeaurora one?
20:20.48cr2we need to finish the hw support first
20:21.15cr2like lcd init/sleep
20:21.19cr2rpc
20:21.22cr2and bt
20:21.39cr2because it's tree-independent
20:21.58cr2wifi
20:25.18*** join/#htc-linux kiozen (n=kiozen@rgnb-5d87c78f.pool.mediaWays.net)
20:26.54dcordes-kaishi
20:27.34dcordes-kaisdruidu any additional thought on the blac/kovs problem?
20:29.49*** join/#htc-linux Untouchab1e (n=Untoucha@82.147.51.146)
20:29.56Untouchab1eso we got audio and mic working now (for the Touch Pro)
20:30.02*** join/#htc-linux Squarc (n=Squarc@82-217-32-29.cable.quicknet.nl)
20:32.15cr2Untouchab1e: mic does not work for me. what about you ?
20:33.33Untouchab1ecr2: I just got back from a military manouver so I havent tested it yet
20:34.14cr2druidu: gpio0x19 is probably the arm9 irq. if it finds the magic number it shutdowns everything
20:34.40cr2dex notify + gpio, it makes a softreset
20:34.53Untouchab1eIs anyone working on Android for the Touch Pro 2?
20:35.09cr2Untouchab1e: on the kernel, not android
20:35.18Untouchab1ecr2 any progress?
20:35.23cr2Untouchab1e: do you have the device ?
20:35.40*** join/#htc-linux Zinbolic (n=zinbolic@84.238.80.215)
20:36.17Untouchab1ecr2 no, I got the RAPH100 and the HTC Hero at the moment
20:36.30Untouchab1ebut the Touch Pro 2 is simply the best looking device out there, both in looks and quality if you ask me
20:37.01cr2you can't develop for a device that you don't have
20:37.22Untouchab1ecr2 I am perfectly aware of that.. I was just wondering if anyone was developing for it
20:37.40cr2actually i don't understand why they can't boot it
20:37.55cr2most things i've seen are the same
20:38.01Untouchab1eThe Touch Pro 2 and Touch Pro is pretty much identical hardware wise as far as Ive read
20:38.04cr2just a different lcd resolution
20:38.11cr2yes
20:38.18AstainHellbringwifi/BT different
20:38.51AstainHellbringheat generation is different
20:39.02cr2lol
20:39.14cr2AstainHellbring: wifi/bt ?
20:39.15Untouchab1ehehe
20:39.18AstainHellbringie tp2 doesnt get hot with excessive wifi tethering like tp did
20:39.25Untouchab1eaha
20:39.36Untouchab1edamnit! Why cant just HTC make a Android version of the Touch Pro 2!!!
20:40.08cr2Untouchab1e: $$$ from m$$$
20:40.14Untouchab1ecr2 Yeah I know :(
20:40.36Untouchab1ebut why cant they make an equally amazing device and let it run Android and Sense
20:40.40druidudcordes-kais: hmm, how about actually printing resource->start, just before `mddi->irq = resource->start`
20:40.50druiduI find it hard to believe that is causing the crash...
20:41.01Untouchab1eCurrently, there are no qwerty android devices even rumoured
20:41.16cr2Untouchab1e: g1 ?
20:41.28Untouchab1ecr2 Yeah, Ive already had that one
20:41.31Untouchab1eugly device
20:41.40cr2yeah
20:41.43Untouchab1eand poorly constructed.. Not very well thought through
20:42.05cr2why doesn't the keyboard work on raph ?
20:42.15Untouchab1esince when did it not work?
20:42.21cr2is the /dev/input/eventX hardcoded somewhere in androed ?
20:42.40cr2on xda-dev they say it's not working
20:42.49cr2check the diamnd thread
20:42.58Untouchab1ethe raph keyboard has been working fine on the packages ive put up on connect-utb
20:43.05*** join/#htc-linux sd00 (n=smac@82-39-194-231.cable.ubr02.jarr.blueyonder.co.uk)
20:43.18Untouchab1eas mentioned, ive been out on a manouver the last few weeks so havent come quite up to date yet
20:43.28*** join/#htc-linux leaigor (n=laigor@188.134.36.14)
20:43.46cr2ok
20:43.58Untouchab1ebut it was working before at least
20:44.14*** join/#htc-linux dwaradzyn (n=chatzill@chello089079067084.chello.pl)
20:44.17dwaradzynhi there
20:44.28Untouchab1ehey dwaradzyn
20:45.06dwaradzynabout the hang during boot on kovsky/blac, i followed what dcordes did today and i have ramconsole dumped
20:45.08dwaradzynhttp://pastebin.com/m1652aa87
20:46.21cr2dwaradzyn: pll0=245 that's making me happy
20:47.31cr2<3>mddi_wait_interrupt 1, timeout waiting for 20000, INT = 21c000, STAT = 8084f2 gotint = c000
20:48.02cr2[118531.628967] <3>mddi_init, timeout waiting for caps
20:48.04cr2[118531.629028] mddi: no client found
20:48.05cr2[118531.629058] mddi powerdown: stat 8085f2
20:48.17cr2epson not found ?
20:48.18druidudwaradzyn: i2c doesn't seem to work
20:48.25*** join/#htc-linux MethoS- (n=clemens@host-091-096-212-099.ewe-ip-backbone.de)
20:48.44dcordes-kaisdwaradzyn this makes me suspect the error is on my side
20:49.06dcordes-kaisdwaradzyn kovs blac have same mem setup, no?
20:49.48cr2but the machine booted
20:49.59dwaradzyndcordes-kais: similar - sure the have, the same - not so sure
20:50.00cr2it's some bug in the mddi init code.
20:50.13*** join/#htc-linux MethoS-- (n=clemens@dyndsl-085-016-164-127.ewe-ip-backbone.de)
20:50.52dcordes-kaisdwaradzyn so you get any xtra output in the conmsole dmp?
20:52.53dcordes-kaisdwaradzyn did you use same ram console patch as me?
20:53.21dwaradzyndcordes-kais: i had no druidu's console enabled in .config. it vibrated twice, i waited a minute (still wm+haret on device lcd), reboot, wifi, dump.
20:53.45dwaradzyndcordes-kais: the same patch, it worked out of the box
20:54.16dcordes-kaishmm
20:54.30dcordes-kaisdid same procedure, same patch
20:54.44dcordes-kaisso kovsky has evil overwrite
20:57.57xperiadwaradzyn> it vibrated twice, i waited a minute (still wm+haret on device lcd),
20:57.59xperiahave the same bug. i cant see how linux is booting !
20:58.07dcordes-kaisoh
20:58.12dcordes-kaisI see it
20:58.28cr2mddi clock too high ?
20:58.37dcordes-kaissis you guys disbale the sapphire config?
20:58.42cr2try to change it back to 0xa00
20:58.51druiduit won't compile without disabling it anyway
20:58.55druiducheck out git
20:59.03dcordes-kaisyea I am so stupid sorry
20:59.08druiduI made a few changes, updated htcraphael_defconfig
20:59.10druiduread the comments
21:00.09xperiadcordes-kais: the problem on kovsky is probably that we cant reboot the device while the blackstone people can reboot. thats why they have the linux dmesg and we not probably. as i know you needed to do a hard reset to get back to winmo from the linux hang.
21:00.16*** join/#htc-linux Doikor (n=aki@192.58.52.211)
21:00.51xperiacr2 where is this 0xa00 ?
21:00.51dcordes-kaiswrong
21:01.01cr2in clock_wince.c
21:01.07cr2PMDH_CLK
21:01.27xperiaokay will checkit. thanks for the helpfull answer
21:01.36cr2druidu: there is no sharp spi table ?
21:01.52dcordes-kaisbrb
21:02.34cr2druidu: hehe. don't see sharp at all in my copy
21:02.53cr2hm. sorry, founf it
21:03.37cr2druidu: you have it much easier
21:03.52cr212,1
21:04.02cr2msleep 12c
21:04.08cr213,3
21:04.15cr23f,0
21:04.22cr2msleep 30
21:04.29cr2that's all
21:04.48dwaradzynxperia: you can enable boot console in menuconfig (sys type -> boot console), you will see messages even before the vibrates
21:05.22druidushould make that console self-contained, remove dependencies on mdp driver
21:05.30cr2ok, back to adsp
21:07.26xperiacr2:
21:07.28xperiaOLD = { .clk_id = PMDH_CLK, .offset = 0x8c, .ns_only = 0xa04, .name="PMDH_CLK"},
21:07.29xperiaNEW = { .clk_id = PMDH_CLK, .offset = 0x8c, .ns_only = 0xa00, .name="PMDH_CLK"},
21:07.31xperiaright?
21:08.07*** join/#htc-linux MethoS_ (n=clemens@dyndsl-085-016-167-098.ewe-ip-backbone.de)
21:08.36cr2xperia: yes, try it
21:09.18xperiaokay rebuilding the kernel. will report in 15 minutes back !
21:09.49cr2druidu: ok, i think now i got it. we need to change some things then.
21:10.09druiduon what? adsp?
21:14.34*** join/#htc-linux cr2_ (n=cr2@ip-77-25-246-243.web.vodafone.de)
21:15.33cr2_damn dhcp
21:16.17druiduyou got it, what? :P
21:16.49cr2_[23:10] <cr2> there are 3 options
21:16.51cr2_[23:10] <cr2> softreboot, shutdown and poweroff
21:16.52cr2_[23:11] <cr2> we do shutdown in poweroff (dex 0x14)
21:16.54cr2_[23:11] <cr2> the poweroff should be magic number +gpio
21:16.55cr2_[23:12] <cr2> softreboot is arm9 notify (commented out now) + gpio
21:16.57cr2_[23:12] <cr2> don't know how it fits into the googel kernel though
21:17.34*** join/#htc-linux MethoS- (n=clemens@dyndsl-085-016-163-077.ewe-ip-backbone.de)
21:19.16*** join/#htc-linux cr2 (n=cr2@ip-77-25-246-243.web.vodafone.de)
21:19.43*** join/#htc-linux Dunedan (n=dunedan@phoenitydawn.de)
21:19.48dcordes-kaisawkward
21:20.49dcordes-kaisxperia I wish you the best of luck fixing the problem.
21:21.55cr2what is awkward ?
21:22.10*** join/#htc-linux Captnoord (n=Captnoor@dc5147a47b.adsl.wanadoo.nl)
21:22.54Captnoordcr2 nice progress :D
21:23.47cr2Captnoord: you have working sound ?
21:23.53CaptnoordI can check
21:23.57Captnoordneed to reboot to build
21:23.59Captnoordbrb
21:25.33*** join/#htc-linux darkstar62 (n=darkstar@97-126-104-157.tukw.qwest.net)
21:26.37*** join/#htc-linux captnoord (n=captnoor@dc5147a47b.adsl.wanadoo.nl)
21:26.43captnoordre
21:26.44xperiadcordes-kais: thanks. just copying the files to the phone.
21:27.10captnoordfetching changed
21:27.21captnoordcrosses his fingers
21:27.22captnoord:P
21:27.30captnoordbtw do I need the playsound file?
21:27.50*** join/#htc-linux MethoS-- (n=clemens@host-091-096-212-137.ewe-ip-backbone.de)
21:28.06cr2it's in the ramdisk
21:28.14cr2playsound2
21:28.41captnoordlemme first build
21:30.36cr2hmm. cdma adsp is more difficult
21:30.54captnoordbuild finished... i'm using a connect-utb.com android package... and I've updated the zImage.... does that package contain the playsound2 file?
21:30.56cr2it has 5 translations tabls instead of 3 for gsm
21:31.13captnoordcr2 more of the same idea ... or different
21:31.30captnoordlemme boot
21:31.31xperiacr2: it makes two vibras. then the screen goes white ! before this patch it makes two times vibra and then the screen frezze
21:32.14cr2xperia: ok, then it's something else
21:32.21captnoordxperia: your sure your using the latest build?.....
21:32.28xperiayes
21:32.37captnoordhmmm... and your commandline...
21:32.42captnoordharet's
21:32.49|Jason8|Raph|xperia: which phone do you have?
21:32.51cr2captnoord: it's a different device
21:32.59captnoordthen I shut up
21:33.14cr2|Jason8|Raph|: do you have mic working ?
21:33.16xperiaJason8: kovsky
21:33.24|Jason8|Raph|cr2: nah.
21:33.50dwaradzynwell, changing PMDH_CLK to 0xa00 make it working for blac100
21:34.12*** join/#htc-linux MethoS_ (n=clemens@dyndsl-085-016-167-231.ewe-ip-backbone.de)
21:34.14|Jason8|Raph|xperia: sounds like you have a wrong zImage and your default.txt is off
21:34.34captnoordI can't remember how I fixed the the problem that it didn't found the sd card...
21:35.01|Jason8|Raph|use the old zImage that had the screen freeze and change your default.txt
21:35.50xperiaJason8: i have booted till yet with out any problems with haret and the default.txt. in fact changing the kernel with the old one works. only the new kernel makes problems.
21:36.21|Jason8|Raph|the kernel that came out two days ago doesn't work on my raphael, either.
21:36.26captnoordit tells me that sound is on... and I didn't had that option before
21:37.13xperiadcordes-kais: on my phone i cant boot the new kernel. please try you with this patch if you can boot linux.
21:38.12cr2dwaradzyn: probably the epson can't process so fast
21:38.22captnoordre-inserting my sd card makes it freeze
21:38.32cr2dwaradzyn: so the right LCD init is pending for you.
21:39.03cr2dwaradzyn: you may try 0xa0c, 0xa41
21:39.33cr2dwaradzyn: hmm. i think your value is 0xa19, like on raph100
21:39.56cr2dwaradzyn: so 0xa0c or 0xa19
21:40.23dwaradzyncr2: i can try, i'm not sure what these values are...
21:40.29xperiacr2: are the values realy for dwaradzyn or for me and decordes. he has written that for him 0xa00 worked.
21:42.10cr2xperia: 0xa00 is the slowest value
21:42.26cr2dwaradzyn: it's the mddi clock speed
21:43.38cr2dwaradzyn: ok, so can you play sound now ?
21:47.13dwaradzyncr2: not tested yet. i'm thinking to boot android now and see (and hopefully hear) what happen. or should i test it another way?
21:50.14xperiahave to go. see you tomorrow cracks. bye !
21:50.18cr2does not matter
21:50.43captnoordI think I was using a old android package....
21:50.45captnoordhmmm
21:50.46captnoordlol
21:51.09cr2druidu: cdma adsp is more difficult than i have thought
21:51.37captnoordcr2 you have a dll dump?
21:51.40captnoordsystem dll
21:51.45captnoordthen I can bootup my ida
21:51.53captnoordand maybe look at it with diff eyes
21:53.14*** join/#htc-linux MethoS- (n=clemens@host-091-096-209-011.ewe-ip-backbone.de)
21:56.51dwaradzynaudio does not work, dialer did not produce any sound (i think it should produce dtmf tones when pressing numbers). during call speaker does not work neither does the mic
21:57.12druiduanalyzing asm code usually goes really slow progress...
21:58.15cr2<PROTECTED>
21:58.25cr2does not look hopeless
21:58.53cr2i'll add 2 fake images, and somebody should just look at the debug output
21:59.26cr2<PROTECTED>
21:59.38cr2this must be bumped by 2 i guess
22:00.16cr2so some testing is needed.
22:01.11captnoordno cdma phone here
22:02.22cr2tmzt: ping
22:02.54cr2druidu: what are we gong to do now ?
22:03.22captnoordtry to take over the world?
22:03.52captnoordis usualy a good thing....
22:03.54captnoordpgu?
22:04.09captnoordbattery saving thingy's?
22:04.24captnoordgpu*
22:04.58cr2gpu works afaik
22:05.05captnoordk
22:05.20cr2closed source in androiid
22:05.23druidudoesn't for me, probably some stuff about resources
22:05.39cr2pmem ?
22:06.56cr2druidu: i see a big revamp of clock-wince.c coming
22:07.05druiduactually pmem is added, don't know why it's not working
22:07.10cr2maybe we should add
22:07.13druiduwas looking exactly at clocks actually
22:07.24druiduwe should adopt structure from aurora
22:07.30druiduit's really convinient
22:07.35cr2static int get_pll_hz(int n) to it
22:07.49cr2to avoid the clock-msm7x30.c complexity
22:08.49druiduso we need a way to determine the best flags to use to get a frequency as close as possible to the requested one, right?
22:09.17cr2if the pll settings remain constant, then it's not a problem
22:10.50cr2we may do static int get_pll_hz(int n), and compare it with the hardcoded values
22:10.58*** join/#htc-linux MethoS-- (n=clemens@dyndsl-085-016-162-077.ewe-ip-backbone.de)
22:11.05captnoordcr2 can you tell me how I can reach playsound2?
22:11.13captnoordtrough terminal?
22:11.21cr2and still use the hardcoded values for the source.
22:11.29cr2captnoord: ssh over usb
22:11.42captnoordk
22:12.00druidunot sure I know what you mean by get_pll_hz
22:12.25cr2druidu: the pll0,1,2,3 values that you print out now.
22:12.53cr2druidu: i want to be sure that we use the right clock for BT
22:14.51cr2druidu: btw, do you have the adsp logs ?
22:15.20cr2druidu: i'm interested which 'images' are used in real life
22:18.18druiduadsp logs?
22:18.37druiduI have a full dmesg log, but only boot, no calls in it
22:18.47druiduI posted the log on pastebin
22:20.20cr2my log was spammed by rpc debug
22:20.51cr2but gps needs to be debugged before we can switch it off
22:21.19cr2ok, it's not really that important
22:23.04cr2druidu: we may calculate the "real" frequencies if the get_pll_hz() is implemented.
22:23.13cr2and log them.
22:23.28cr2i have some concerns about the PLL0 based clocks
22:23.43cr2and want to log the calls
22:24.23cr2the PLL1 is always 768MHz ans PLL2 is used only by the arm11 itself.
22:25.15cr2bt clocks are pll0 based . in wince.
22:25.39cr2spl uses some hacked pll1 based frequency close to 115200 :)
22:29.22druiducr2: start syslog daemon in /init
22:29.28druidusave messages in the ram filesystem
22:30.10druiduwhat exactly is TCX0?
22:31.57cr219.2MHz xtal
22:32.52*** join/#htc-linux MrPippy (n=pip@adsl-75-11-160-45.dsl.sndg02.sbcglobal.net)
22:33.00cr2the same on all msm chipsets
22:33.53cr2druidu: the clock math (for MD/NS) is relatively primitive
22:34.24cr2frequency=S*M/N/P
22:35.57cr2S is source clock frequency. it may be TCX0 (if =0), PLL0(if =4), PLL1 (if =1). don't know about the others, but we don't need it i think
22:36.18cr2M is known, N is known, and P is ...
22:36.36cr2let me check the s? values. need to update wiki on that ;)
22:37.15cr2#define MSM_CLOCK_REG(frequency,a1,a2,a3,a4,a5,a6,a7)
22:37.44cr2a1=M
22:37.57cr2a2=N
22:38.38cr2a3=D (usually D=N/2, or sometimes N/2-1)
22:39.02cr2D is needed only for PWM.
22:39.12*** join/#htc-linux scatterp (n=deminish@94-170-84-26.cable.ubr16.enfi.blueyonder.co.uk)
22:39.29scatterphi can the G1 run linux instead of android ?
22:40.05scatterpif not whats the most powerful htc with a keyboard that runs linux well..
22:40.15kri5scatterp: what do you mean by linux? because G1 already run on linux kernel with android
22:40.27cr2P=a4+1
22:41.01ali12341scatterp: does it have to be htc because the n900 looks like it will be a lot better than anything htc has released yet...
22:41.16cr2S=a6=s2
22:41.20scatterperm something i can run standard linux apps on
22:41.28scatterpsuch as aircrack for example
22:41.30cr2scatterp: i think g1 can run debian
22:41.42scatterpcr2 debian would be perfect
22:41.55kri5scatterp: so an standard GNU/Linux distib?
22:41.59mickeylit can run debian, but don't expect all the hardware to work
22:42.03scatterpofc
22:42.26scatterpim also wondering about the bluetooth chip its got inside
22:42.31scatterphciconfig hci0 revision
22:42.33mickeyln900 will also be half-closed
22:42.33scatterpoutput
22:42.38mickeylso don't have too high hopes
22:42.39cr2druidu: so it's #define MSM_CLOCK_REG(frequency,a1,a2,a3,a4,a5,a6,a7) #define MSM_CLOCK_REG(frequency,M,N,D,P-1,a5,S,1)
22:43.04cr2scatterp: brf6300
22:43.11ali12341yes, nokia stuff has a lot of closed binaries too. but n900 CPU is just way more powerful than any htc phone
22:43.16cr2scatterp: i think it's 2.0
22:43.32scatterpis that a rom or can it be reflashed
22:43.36scatterpto support sniffing
22:43.44mickeylyeah, if that's "open enough" for you, ok
22:43.48cr2ali12341: powerful = needs more watts ?
22:44.06ali12341cr2: true, but nokia PM usually works quite well
22:44.37scatterpcr2 would the bluetooth be Chip version: BlueCore4-External
22:45.04cr2mickeyl: i can live without a cam, and 3d. everything else will work.
22:45.13mickeylon the N900?
22:45.14cr2scatterp: no, it's BRF from TI
22:45.19cr2mickeyl: on htc
22:45.23scatterpok
22:45.36mickeylcr2: you're an optimist.
22:45.38mickeylbut that's ok :)
22:45.52mickeyli don't care about cam and 3d either
22:45.58cr2mickeyl: cam is just isn't worth the pain to RE
22:46.12ali12341phone cameras are usually pretty bad anyway
22:46.22cr2and openGL is too much work.
22:46.56ali12341this irq bug is really annoying me
22:47.11ali12341i fixed a bug that totally screwed up the irqs, but it still doesnt work
22:47.25ali12341the bug looked like the cause, but wasn't. so annoying
22:47.38cr2ali12341: we get a flood of BT irqs here. so i feel your pain ;)
22:47.53ali12341we get NO irqs at all
22:48.22cr2that's weird.
22:48.35ali12341it's like they are globally masked, but they arn't, and the code to enable them hasn;t changed since the working version
22:48.45cr2ali12341: at least you have some docs for your cpu
22:49.22cr2and we have some shitty google driver, that declares RX irq, but is not using it.
22:49.51mickeylboots HEAD on raph
22:50.00cr2hehe
22:57.16*** join/#htc-linux miknix (n=miknix@gentoo/developer/miknix)
23:01.15MrPippyusb (for usbnet) isn't working for me with latest version, any ideas why?
23:02.37*** part/#htc-linux iseeka (i=5aa86195@gateway/web/freenode/x-hnjgmboydqqbifbm)
23:09.27mickeylsame here on raph100
23:10.43mickeylah, it's just the interface not coming up on its own
23:11.20MrPippyhmm i'll try that
23:11.21mickeylallow-hotplug usb0 doesn't seem to cut it yet
23:11.42mickeyli guess we're still missing that usb pull-up irq or something
23:14.02mickeylinitial status of backlight logic still wrong btw.
23:14.06mickeylerr
23:14.13mickeyls/backlight/kbd-led
23:14.16mickeylreports 0 while being 1
23:14.18cr2mickeyl: usb works for me
23:14.35MrPippyhmm my init has a 'busybox ifconfig usb0 up' in it, it all worked in older kernels
23:14.37mickeylyes, here as well, just doesn't come up by default. must be userland problem
23:15.01cr2works with my initd
23:15.28mickeylok, so which new features do we have that i can test/support?
23:15.35mickeyl(compared with a month ago)
23:16.14cr2playing sound
23:17.00mickeylok. let me look for your test program
23:23.12mickeylcan't find it in the wiki
23:24.09mickeylhttp://linuxtogo.org/~lgorris/misc/playwav2.c ?
23:25.45scatterpcr2 you mentioned that debian is semi-perfect to you know of anything specifically that does not work from reading up on it... it would seem that there are no issues.. not that i could find mentioned anywhere anyway
23:31.51mickeylwah, that's a brutal volume :)
23:32.00mickeylbut indeed, i hear something
23:32.03mickeylcongrats
23:33.01cr2change the volume value
23:34.09mickeylin the sourceß
23:34.09mickeyl?
23:34.46cr2yes
23:34.50mickeylk
23:36.18|Jason8|Raph|Hey guys, anyone play with that supposed clock-setting command for Haret/Android?
23:36.45*** join/#htc-linux rafyvitto (n=rafyvitt@24.54.253.34)
23:37.58cr2good night
23:42.08mickeylhmm, tweaking volume doesn't work
23:42.22mickeylwhat about the mp3 device, anyone got that to work yet?
23:44.41|Jason8|Raph|this android stuff is freakin' awesome.
23:45.14mickeylhmm
23:46.18tmztit appears windows mobile doesn't use it
23:46.23tmztjust pcm
23:46.24*** join/#htc-linux ellisway (n=ellis@94.30.45.128)
23:47.08mickeylk, the mp3 device is there, but can't be opened
23:47.39mickeyldo these ioctls also control routing? like, what do i need to do to connect BP audio to mic/earpiece
23:47.41mickeyl?
23:51.32mickeyl'night
23:56.30tmztcr2: what images?
23:56.43tmztyou have zImage for me to test?
23:57.40*** join/#htc-linux xfiles (n=xfiles@c-98-193-151-150.hsd1.ky.comcast.net)
23:58.08*** part/#htc-linux xfiles (n=xfiles@c-98-193-151-150.hsd1.ky.comcast.net)
23:59.11tmzttranslations table?

Generated by irclog2html.pl Modified by Tim Riker to work with infobot.