Note: This logging service will stop october 2014. If someone fancies taking it over, contact me.


<2010>  <January>  Show everything  Back to index      

00:00 phh cr2: I implemented h2w
00:00 phh and headset feature works
00:00 stinebd sigh
00:00 phh stinebd: ?
00:00 stinebd after roughly 75 minutes, the phone has apparently finished charging and is going into deep sleep periodically while on usb
00:00 phh stinebd: yes, vbus detection sucks.
00:00 NetRipper cr2, ah ok
00:00 NetRipper thx
00:00 stinebd but the clock is still right
00:00 phh chamonix: have fun with stinebd.
00:01 chamonix hhuh?
00:01 chamonix awww
00:02 phh cr2: I still get an i2c error :/
00:02 phh oops
00:02 phh one bad copy paste.
00:02 stinebd am i doing it wrong?
00:02 phh cr2: it was gp1/gp3/gp5 ?
00:02 chamonix well let's do some testing then... I remember vaguely we were at sleep_mode=3 back then
00:12 cr2 phh: vbus detection is readl(smem+0xxxC)
00:12 cr2 phh: h2w needs gpio alt to uat1/uart3 (depending on the phone) switch
00:12 cr2 phh: and some power magic
00:16 phh yes it is.
00:16 phh stinebd: I don't think so :p
00:16 phh have fun :p
00:16 stinebd and no idle_sleep_mode right?
00:16 phh stinebd: it's this one chamonix is talking about
00:18 phh i guess
00:23 cr2_ [00:03] <cr2> phh: vbus detection is readl(smem+0xxxC)
00:23 cr2_ [00:04] <cr2> phh: h2w needs gpio alt to uat1/uart3 (depending on the phone) switch
00:23 cr2_ [00:04] <cr2> phh: and some power magic
00:23 cr2_ phh: maybe !readl()
00:23 phh cr2: groumpf with gp1=1800;gp3=2850;gp5=2650;gpio RAPH100_CAM1_RST set to 0 then 1, I still get no I2C answer
00:23 stinebd ?
00:25 phh argh wrong i2c addr
00:25 cr2_ phh: other cam gpios state ?
00:25 stinebd synced to the second :o
00:25 stinebd this thing keeps better time than my laptop
00:29 cr2_ phh: are you doing other gpios right ? and mi380 mux gpio.
00:29 cr2_ lol
00:29 cr2_ 6c/2=0x36
00:29 cr2_ afair it's the same on g1
00:30 phh cr2: is it base battery +22*4 ?
00:30 phh (22 decimal)
00:30 phh [ 15.579589] msm_gpio_set_function: could not set alt func 7 => 1
00:30 phh uh ?
00:30 chamonix phh: I just read an interesting commit on android git about loc
00:30 chamonix Fix bug 1687659 - cell id in reference ril.
00:30 chamonix The LAC and CID of RIL_REQUEST_REGISTRATION_STATE should be in the hex format. The valid LAC are from 0x0000 to 0xffff and Valid CID are from 0x00000000 to 0xffffffff. reference-ril should return hex value instead of decimal one.
00:30 cr2_ lol
00:30 phh [00:18:56] <phh> cr2: is it base battery +22*4 ?
00:30 phh [00:19:44] <phh> (22 decimal)
00:30 ToAsTcfh hey atleast im on the big half of the split
00:31 chamonix I'm gonna try replacing the format of all those cid+lac returns
00:31 phh cr2: in the cam init sequence, which gpio are triggered ?
00:31 phh the driver I have only does reset 0 then 1
00:31 phh not even 0 actually
00:31 phh yes trout uses the same i2c
00:31 phh cr2: which other gpios ?
00:31 phh {2..15} should be ok
00:31 phh but I don't know what I am supposed to do with 1e, 5b/5d/6b/6d
00:31 phh chamonix: any url ?
00:31 camden_ I'm excited about 2d and 3d accelleration.
00:31 cr2_ phh: alt=7 ???
00:32 ToAsTcfh phh: i2c s5k3e2fx?
00:32 phh chamonix: yes we do %d everywhere.
00:32 tmzt glem: I might do it the way ubuntu does, have an */ubuntu modules path for instance
00:32 tmzt but I don't know, does android use modutils?
00:37 chamonix;a=commit;h=6f189a6a1b53adfa9570cf5911c3f81ef1b1fcf8
00:37 chamonix it's an old one but who knows
00:37 phh i'm trying that change now.
00:37 cr2 phh: how do you get alt=7 ?
00:37 tmzt cr2: yes?
00:37 cr2 tmzt: i'm looking at my rhod notes now
00:37 phh chamonix: gg.
00:38 cr2 phh: ?
00:38 phh joee_: please use for the next release
00:38 herpez|sometimes im a bit confused. in Commits in linux-on-qualcomm-s-msm/linux-msm:htc-msm-2.6.27 there arent the commits from glemsom's
00:38 stinebd :o
00:38 phh this fixes celllocalisation thanks to chamonix
00:38 tmzt nice
00:38 tmzt what do you need for rhod500?
00:38 tmzt yeah, defines are not going to be in -panel anymore
00:38 stinebd awesome
00:39 tmzt I don't really like the microp_spi_table, it's not specific at all and shouldn't be microp sepecfic either
00:39 tmzt it's just spi
00:39 tmzt (the name)
00:39 stinebd i'm excited about linux kernel booting on the devices
00:39 stinebd all that high-level stuff i can get excited about later
00:39 phh chamonix: pushed if you didn't do the change yet :p
00:39 phh camden_: me too but I'm fed up of all these api changes.
00:39 phh cr2_: it's gpio7,alt5
00:39 chamonix I've located the 3 functions dealing with cid+lac, I'll switch them to %x
00:39 chamonix I don't see the relation to the gmaps issue as I get the info right from telefony manager in java but it's worth a try as I know how to build/install RIL now :)
00:39 chamonix I've tried different other fixes about neighbour cells as well but to no avail...
00:41 camden_ phh: why not freeze the android version until everything is settled and stable? or has this already happened with 2.1
00:41 chamonix gg?
00:41 cr2 tmzt: did you ever traced i2c on rhod ?
00:41 chamonix phh: don't tell me u tried it already OO
00:41 tmzt somebody understands RILJ?
00:41 camden_ ?
00:41 tmzt the java ril stuff
00:41 tmzt cr2: what does that mean? do you have enough to get asoc routing to work?
00:41 tmzt no
00:42 tmzt I have rhod500, one of the cdma/gsm versions, but I assume that part is mostly the same
00:42 cr2 phh: i'm comparing raph100 cam iwth my my topa notes. they lokk identical.
00:42 cr2 phh: the wiki "alt" value needs to be parsed
00:42 phh hum alt5 doesn't exist anyway
00:42 phh oh it's not alt5 it's func 5 fine
00:42 phh same as reported on the wiki
00:42 phh cr2: it's gpio7 alt5
00:42 phh not alt
00:43 phh herpez|sometimes: glemsom do no commits
00:43 phh cr2: printk(KERN_WARNING "%s: could not set alt func %u => %u\n", __func__, cfg.gpio, cfg.func);
00:43 herpez|sometimes phh - compiling the modules gives me 4 tar files. which one shall i use? (glemsom didnt compiled the last kernel)
00:43 chamonix phh: I'll checkout again since my code is messed up with other changes that didn't fix anything....
00:43 phh the '7' is not the alt, but the gpio
00:43 phh ToAsTcfh: :)
00:43 cr2_ alt5 is for cam pins
00:43 cr2_ you need to parse it
00:43 phh tmzt: it is sent through microp though
00:43 cr2_ phh: like 1c0/1c4 for mclk
00:43 cr2_ phh: how will you set gpio 0xf to alt 1c4 ?
00:44 stinebd good good?
00:44 phh chamonix: -_-' read your highlights.
00:44 phh chamonix: it works, and I already pushed it.
00:44 phh camden_: the more we freeze the more we will have to redo.
00:44 phh anyway flan should be the same as eclair but the GUI afaik
00:44 phh chamonix: bah what ?
00:44 phh it takes less than 1 minute to do
00:44 chamonix omg
00:44 tmzt phh: right, and someday that will be a mfd/spi host
00:44 chamonix \o/
00:44 phh chamonix: I read and understood the whole RIL stuff, so I know where I have to look :p
00:44 camden phh: do they really release that often?
00:44 phh cr2_: the function is the 2 lower bits afaik,so this makes a function=1.
00:44 stinebd yes it does indeed work
00:44 stinebd i'm trying it now
00:44 tmzt so what's this function stuff?
00:45 phh stinebd: what ?
00:45 phh he didn't ? damn it means I broke something
00:45 phh again
00:45 stinebd chamonix: works as expected, nice job
00:45 chamonix see phh that's the difference btwn you and me.... sometimes I have the feeling I understand something but most of the time I'm more confused after browsing the code than before ;-)
00:45 tmzt still gpio or something with shadow registers?
00:45 tmzt or microp?
00:45 phh stinebd: it works as expected for everyone but him :p
00:46 stinebd phh: that's ok as long as it works for us ;)
00:46 phh cr2_: read trout's code ? :D
00:46 chamonix lmfao
00:46 cr2_ phh: you need to set the names
00:46 phh cr2_: I haven't exactly the same thing actually, I get 1c0
00:46 phh camden: they break everything quite often.
00:46 camden that sucks
00:46 tmzt gconf recursive-unset :)
00:46 cr2_ phh: trout code is braindead, we are forced to use this PITA interface.
00:46 phh chamonix: can you see the big function which redirects every call to some functions ?
00:46 chamonix btw stinebd, home++ is somehow wierd but still interesting... do you know if beta is only something to fix users preparing a paid version or is it real free software?
00:46 cr2_ phh: 1c4 is when it's enabled
00:46 MrPippy wow wow wow i'm running with bank2 and its stable
00:47 chamonix yes phh, I see it even if I close my eyes
00:47 cr2_ MrPippy: with this recent patch ?
00:47 phh chamonix: for donut (or was it eclair ?mmmm), I had to rework it from 0, by reading the ril header
00:47 phh that helps a bit :p
00:48 phh MrPippy: with the +72MB ?
00:48 MrPippy yeah, with patch and +77MB
00:48 phh huhu
00:48 chamonix that's reassuring.... I mean I know every part of my motorbike too after having taken it apart and remounted it :)
00:48 cr2_ phh: 1c0 is on cam poweroff
00:49 phh MrPippy: commit the patch and set extra=1 in pmem.c then
00:49 tmzt cr2_: I'm a little lost here, what's 1c0?
00:49 phh cr2_: oh really ?
00:49 phh tmzt: gpio function for gpio .xf
00:49 phh 0xf*
00:49 cr2_ tmzt: cam_mclk alt gpio
00:50 chamonix phh: while reading RIL I was wondering if the SIM entry is exclusively handeled there and "only" missing or if there are other depends I didn't get yet
00:51 tmzt offset?
00:51 phh chamonix: now it no longer works /o\
00:51 cr2_ phh: at least for mi380
00:51 stinebd chamonix: don't know for sure, but i guess it will be paid
00:51 cr2_ phh: well, for main cam too
00:55 MrPippy haha i'm playing music with pandora and its stable and working!
00:55 emwe MrPippy: for real with bank2?! i spent days getting a hold of that stuff, trying different pmem setups ...
00:55 phh it thinks i'm in paris.
00:55 phh tmzt: it's just a gpio config value
00:55 phh tmzt: you can set many things for each gpio
00:55 cr2_ tmzt: alt value to be written into the reg
00:56 emwe trying bank2 before or after pmem ....
00:56 MrPippy yeah i spent longer than that with pmem setups, but it was a real kernel bug/limitation
00:56 stinebd :(
00:56 cr2_ phh: i'll do an awk parser
00:56 cr2_ phh: i had one in perl, but it's lost somewhere
00:56 tmzt okay
00:56 phh cr2_: really ? sounds weird to me ... anyway, I have to go to sleep.
00:56 cr2_ yes
00:56 tmzt cr2_: phh: there's one in the omap tree that exposes through debugfs
00:56 tmzt I would really like to see that made generic, possibly with the method used in haretconsole
00:56 tmzt (onebits/twobits)
00:57 cr2_ tmzt: omap ?
00:57 tmzt omap.git, but it's from ti I think
00:57 tmzt it exposes gpios and altfunctions (pinmux) through dbgfs
00:57 emwe MrPippy: i setup some spreadsheet in order to not loose track ;)
00:57 tmzt for dumping the ce configuration it could be useful
00:57 MrPippy emwe: what device are you on?
00:57 emwe MrPippy: topa...
00:58 emwe i had it once going with bank2 MSM_EBIN_BASE + 64MB and afterwards pmem
00:58 emwe but i think that was with radio off
00:58 emwe even worked. then phh pushed and people reported crashes... so just a good boot i guess ;)
00:58 chamonix phh: is the response at least steady or do you still get the "temporarily unavailable" error.... looking at logcat -b radio I noticed that requestNeighboringCellIds sometimes fails and that may be the reason y it returns you to some default place.... for me it's berlin
00:58 chamonix once I was in island :)
00:58 tmzt what's EBIN?
00:58 cr2_ tmzt: should be EBI1
00:58 emwe EBI2, no? ;)
00:59 cr2_ phh: 0x1e and 0x6d are used to select the cam
00:59 tmzt okay, good
00:59 emwe at least that's the define in pmem.c module...
01:02 tmzt mickey|office: get anywhere with fso and asoc?
01:02 cr2_ emwe: EBI2 is nand
01:03 tmzt the ram part of the nand chip?
01:05 herpez|sometimes while making make ARCH=arm CROSS_COMPILE=arm-none-linux-gnueabi- zImage
01:05 herpez|sometimes i get make: arm-none-linux-gnueabi-gcc: Command not found
01:05 herpez|sometimes strange is that this command worked 1 week ago :s
01:05 chamonix u r missing the bin of the crosscompiler in your path maybe?
01:05 cr2_ tmzt: physically
01:06 cr2_ tmzt: but not on the hw interface
01:06 cr2_ tmzt: nand is too slow, so there is EBI2 for it
01:06 tmzt just use the full path, or set PATH on the command line
01:07 tmzt what does it do?
01:09 makkonen hmm. getting a call while eclair is booting is not a great experience. vibrate seems to be stuck on now. and the device is hung.
01:10 Boydell stick it down your pants ;) jk lol
01:11 chamonix wow, I hadn't noticed there's a feature to power off RAPH in the latest Xandroid
01:11 makkonen i said the DEVICE is hung. get your mind out of the gutter.
01:11 ToAsTcfh phh: so are u guys using msm_camera or did u fix up libcamera to work with the other?
01:11 makkonen (I don't even know what that means)
01:12 Boydell a feature to power off RAPH?
01:12 chamonix lol, I should read release notes :-)
01:12 chamonix slaps himself
01:12 stinebd does it work for you?
01:12 chamonix let me see
01:13 herpez|sometimes solved using full path
01:13 tmzt Boydell: init 0?
01:14 emwe cr2: thanks for the info.
01:14 chamonix almost... it hangs at "shutting down" after having turned off radio
01:14 tmzt poweroff?
01:14 tmzt chamonix: what's Xandroid?
01:14 herpez|sometimes and i guess sometimes its needed a new shell to "apply" the path
01:14 chamonix 15.01 tmzt
01:14 stinebd chamonix: same for me, no difference from previous releases
01:14 tmzt herpez|sometimes: source it
01:14 chamonix check release notes
01:14 tmzt source .bashrc
01:14 emwe MrPippy: are you going to push your changes soon? are you gonna use the whole bank?
01:14 tmzt what is Xandroid?
01:14 makkonen power off finally works for me with the latest builds.
01:14 ToAsTcfh or anyone what camera r u trying to get to work earlier
01:14 makkonen tmzt: xdandroid, babijoee's eclair build
01:15 stinebd makkonen: cdma right?
01:15 herpez|sometimes tmzt: xandroid is the name choosed to babihooes build
01:15 makkonen stinebd: yes.
01:15 chamonix lol stinebd, I didn't notice since I skipped updating for one whole week
01:15 stinebd MrPippy pushed a shutdown fix, but for cdma only. babijoee said it's working for him, though.
01:15 MrPippy emwe: i'll push soon, right now i'm testing with 77MB from bank2 and the last 51MB for pmem
01:16 MrPippy stinebd: does everyone want that fix? i can take out the if thats keeping it on diamond
01:16 emwe MrPippy: i got 64MB + 51MB working. but that was pure luck it seems. isn't there said to be some radio basband stuff somewhere in the banks whcih sould not be "touched"?
01:17 emwe MrPippy: I had set up a little spreadsheet. dunno if it might be of use for you... little tables where you could set offset sizes and see the mem ranges taken up by which component...
01:17 MrPippy now i have 107MB + 77MB
01:17 emwe bah. great MrPippy!
01:17 MrPippy but we've always been able to use the 107MB in bank1, not sure if this will help you use more of your bank1
01:18 stinebd MrPippy: i have no idea actually
01:18 stinebd MrPippy: i'll try it locally and let you know
01:18 emwe MrPippy: i could use 108... but anything after screws. seems gpu0 must be @ 107+8MB
01:19 emwe MrPippy: I even tried reloacting gpu0 into MSM_EBIN_BASE range and then let bank0 claim a bit more after 107MB but that did not work
01:19 emwe but GPU0 in MSM_EBIN_BASE did work, though...
01:19 MrPippy yeah we've usually had gpu1 at the end of bank1 and gpu0 in smi
01:20 emwe oh ah sorry... mixed that up... GPU1 can be reloacted to "EBIN". that worked. but you couldn't claim it's "ususal" place in EBI_BASE+107 with bank0
01:22 cr2_ 0x73= ?
01:22 cr2_ bank1 should work with 115MB
01:23 emwe cr2_: and gpu1 afterwards?
01:23 cr2_ nd bank2 of EBI1 is 128MB
01:24 cr2_ you only need need to use alloc_bootmem for the contiguous buffers where appropriate
01:24 emwe cr2_: i think that is out of my scope ;)
01:25 cr2_ ok
01:26 cr2_ good night
01:26 emwe or lets rephrase that ... i would have to dig into regions i don't have a clue of, yet.
01:26 emwe eek, to late ,)
01:28 MrPippy ok pushed
01:29 MrPippy i haven't changed anything in pmem, so you'll need to enable bank2 either with 'pmem.extra' or just by commenting stuff out ;-)
01:31 emwe building ....
01:34 emwe btw, does anybody know of a quick winmo "switch-usb-state-from-as-to-usb-and-vice-versa" app?
01:35 Boydell no, but man, that would be handy!
01:35 emwe happy that commmgr has a menu with usb-setting in HolyGrail R6 but that's still tedious if you don't want to unplug/plug the cable for getting the sd-card back...
01:35 emwe Boydell: indeed ;)
01:36 MrPippy the usb-to-pc settings panel can do it without having to unplug the cable
01:36 Boydell I never look at the microsoft market thing. Never even knew it existed until a little while ago
01:36 tmzt neither does anyone else
01:36 emwe yah, but it's tedious reaching and switching. a small app when launched, detects state, and switches...
01:38 joee_ is the glemsom autobuild working?
01:39 emwe MrPippy: hm, are you testing with pmem.extra=1 or have you commented the bank1 #if 0? didn't work out for me.... i think it didn't work out before either (getting the module param respected)
01:40 MrPippy yeah module params have never worked for me, i'm not sure why
01:41 emwe rebuild ... :)
01:47 emwe hm. froze on switch to android splash...
01:48 MrPippy aww it finally crashed after an hour...huge improvement though
01:49 Boydell MrPippy: this is for your diam500 right?
01:49 MrPippy yeah
01:49 emwe i think i read somewhere that the radio extracts baseband data into the end of EBIN
01:50 emwe trying bank1=64MB now so that there is a little space left...
01:50 Boydell the diam500 and raph800 are similar right? not sure exactly what we have for mem
01:50 camden well, folks, I'm off. have a good one!
01:50 camden thanks again for the working android build!
01:50 emwe night...
01:51 MrPippy yeah real similar
01:56 herpez|sometimes anyone here can help me to compile the modules please?
01:56 emwe MemTotal: 169168 kB!
01:56 emwe ;)
01:56 emwe herpez|sometimes: compcache wlan and compat-msm-wireless?
01:58 emwe MrPippy: i think 64MB might be even too much ... digged up that thread again:
01:58 Boydell have you looked at the wave? good incrustions there
01:59 emwe 19+32MB for baseband stuff.. this might explain the SMI whole at least and the rest might really be expanded into end of EBIN
01:59 herpez|sometimes emwe - i compile the compat-msm-wireless and it makes a tar.gz with just 5 files wl1251.ko mac80211.ko wl1251_sdio.ko cfg80211.ko rfkill_backport.ko
01:59 emwe herpez|sometimes: I've set up a little script which just build everything with correct params set.
02:00 herpez|sometimes however glemson has about 10 or more files
02:00 emwe he has ramswap xvmalloc (both compcache) and wlan.ko (tiwlan)
02:00 emwe s/ramswap/ramzswap
02:00 emwe plus the others...
02:01 herpez|sometimes i also have wlan.ko
02:01 babijoee Glemsom: autobuilds doesn't seem to be working
02:01 herpez|sometimes however i miss compcache
02:01 babijoee Glemsom: please fix :)
02:01 emwe herpez|sometimes: get compcache from google code..
02:01 emwe i used the 0.5.4 version.
02:01 herpez|sometimes ok, thanks
02:01 herpez|sometimes btw, i used this script i made to compile all, if you want to check -
02:01 emwe and then i am copying all modules together and taring them...
02:02 herpez|sometimes emwe - shall i try the 0.6 ver or 0.5.4?
02:02 stinebd 0.5.x
02:03 stinebd 0.6.x won't work with 2.6.27
02:03 emwe i think 0.6.x needs a newer kern, eh?
02:03 emwe ah ;)
02:04 herpez|sometimes the sintax to compile compcache is just make?
02:04 stinebd no
02:04 stinebd it's an external module build, you have to use the kernel build system
02:05 emwe make ARCH=arm CROSS_COMPILE=arm-none-linux-gnueabi- KERNEL_BUILD_PATH=$KERNEL_PATH -j2
02:05 emwe wait .. let ma pastebin mine...
02:05 ToAsTcfh modprobe: chdir(/system/modules/lib/modules): Not a directory
02:05 ToAsTcfh wtf i know its there
02:05 MrPippy emwe: htc's wince and android phones use different amss, those aren't always comparable
02:05 MrPippy you can try that pmem mapping if you want though
02:05 stinebd ToAsTcfh: not a directory?
02:06 stinebd did you touch instead of mkdir? ;)
02:06 emwe
02:06 herpez|sometimes ok, checking
02:06 ToAsTcfh its on my phone
02:07 emwe MrPippy: yes, but the baseband data must be somewhere... or we are lucky and it's at the end of SMI ;)
02:07 ToAsTcfh ug i pushed to the dir
02:07 ToAsTcfh i even went in with astro and confermed its there
02:08 MrPippy i think its usually at the end of ebi bank1, so ebi1: 0-107MB os+applications, 107-115 gpu1, 115-128 amss
02:08 emwe if it is, i am happy with that ;)
02:09 emwe but 77+51 didn't work out for me with radio and wifi enabled.
02:09 emwe and you also had the crash ;)
02:09 emwe testing 64 currently...
02:11 ToAsTcfh stinebd: ur right my bad. when i pushe my module it got all funked up. thanx
02:11 herpez|sometimes emwe - whats the folder generated?
02:11 herpez|sometimes ah, understood
02:12 emwe herpez|sometimes: it's the folder whery i copy the modules and zImage too and where the tgz is kept
02:12 emwe so keeping it a bit clean
02:12 emwe script was inspired by the compat-msm-wireless build skript...
02:12 herpez|sometimes sure
02:12 herpez|sometimes its good, im reading it
02:12 herpez|sometimes i thought we should use arm-2008q3
02:12 herpez|sometimes not 2009
02:12 herpez|sometimes works with 09?
02:12 tmzt q1 I think
02:12 MrPippy emwe: my crash wasn't memory stuff, it was something networking/interrupts
02:12 emwe you might drop some make clean in the scripts...
02:13 emwe MrPippy: ah ok. i see..
02:13 herpez|sometimes tmzt - 2009q1?
02:13 MrPippy i think we're all using 2008q3
02:13 tmzt 08q1
02:13 emwe 2009q3 WorksForMe (c)
02:13 tmzt well..
02:13 herpez|sometimes yes, i'm mypippy, and i thing glemsom is too
02:14 tmzt there was an issue earlier
02:14 tmzt guess it's solved
02:14 emwe ah
02:14 emwe grml... freeze opening the browser ...
02:16 MrPippy whoa yeah i got a crash really fast too, right at the lock screen
02:16 emwe testing 48MB now...
02:16 MrPippy my first test was without hw3d, maybe that makes a difference
02:16 emwe 'hm... but isn't hw3d just using gpu0+gpu1 resources?
02:17 emwe or i am just too stupid reading the code ;)
02:17 MrPippy yeah it is, maybe those aren't in the right place or something
02:18 emwe hey, but android is coming up.
02:18 THE_GFR|WORK where can I get the android release for HTC Touch Pro?
02:18 THE_GFR|WORK or whatever
02:19 emwe when i was tinkering on that earlier it usually crashed in mem_move_free_pages() or something. so this is really great now MrPippy. One just needs patience and a good portion of non-scientific research ;)
02:20 THE_GFR|WORK where can I download the files required?
02:21 emwe THE_GFR|WORK: sorry, no idea. perhaps search on xda?
02:22 emwe oh no! move_freepages_block() crash again...
02:24 THE_GFR|WORK ok
02:24 THE_GFR|WORK thanks
02:29 MrPippy yeah i'm getting panics too
02:38 MrPippy yup move_freepages_block()
02:39 emwe 32MB is booting currently...
02:45 MrPippy hmm i wonder if i need to bring in other patches for this to work right
02:46 emwe 32MB is still running...
02:47 emwe i read that the baseband stuff might expand up to 40mb ... so if it is perhaps really hiding in EBIN too .... 32+51 is 83 ... and the rest might really be hooked up by the baseband data
02:48 emwe no idea really
02:49 MrPippy i'm going to try with one bank now
02:49 emwe one bank aka?
02:51 emwe anyways... damnit... just opened taskkiller and it hangs again...
02:51 emwe and now i am off to bed. good night.
02:54 herpez|sometimes just for the records, i compiled the last kernel commit and blackstone backlight is fixed!
03:40 ivanmmj Can anyone here read chinese?
04:02 MrPippy i'm reverting the memory patch, it just isn't stable at the moment
04:02 MrPippy 15 months passed between our 2.6.27 and this patch, i probably need to backport some other patches too
04:15 makkonen anyone know the correct parameters to pass to mksquashfs?
04:30 babijoee makkonen
04:30 babijoee what are you stuck on?
04:30 stinebd babijoee: did you see phh's message about the new ril?
04:30 babijoee yeah
04:30 stinebd :>
04:30 babijoee i've already incorporated it
04:30 babijoee whats new about it?
04:30 stinebd geolocation
04:30 babijoee ooo
04:31 stinebd chamonix did it
04:31 babijoee ah kool
04:31 babijoee makkonen, -sudo unsquashfs system.sqsh
04:32 babijoee -make changes
04:32 babijoee -sudo mv ./squashfs-root ./system
04:32 babijoee -sudo genext2fs -d ./system -b 100000 -a system.ext2
04:32 babijoee -mkdir ext2mount
04:32 babijoee -sudo mount -o loop ./system.ext2 ./ext2mount
04:32 babijoee -sudo mksquashfs ./ext2mount newsystem.sqsh
04:34 makkonen I don't need to pass any sort of blocksize to mksquashfs?
04:34 stinebd nope
04:34 stinebd and really everything in between "make changes" and "mksquashfs" is optional
04:35 stinebd you can make changes and just squash up the squashfs-root dir
04:36 babijoee 100000 should be enough
04:37 babijoee stinebd: are you sure?
04:37 babijoee stinebd: when i do that the system loops
04:37 stinebd loops how?
04:37 babijoee like the nexus screen loops forever
04:37 stinebd as long as you keep the perms correct it should go fine
04:38 babijoee hmmm not to sure stinebd
04:38 stinebd i've done it
04:38 stinebd quite a lot
04:38 babijoee hrmmm
04:38 stinebd all that testing with the damn key layouts
04:38 stinebd ;)
04:39 babijoee how do you build the sqsh?
04:39 stinebd sudo mksquashfs-3.1 squashfs-root system.sqsh
04:39 babijoee if you dont include the last step?
04:39 babijoee ok
04:39 babijoee fair enough
04:39 stinebd (the -3.1 is a gentoo thing)
04:40 babijoee well thats how i do it plus its in a script
04:40 babijoee so no biggie
04:40 babijoee stinebd: glemsom autobuild is down :(
04:40 stinebd is it?
04:41 stinebd i don't use those
04:41 babijoee ah k
04:41 babijoee its not building new ones
04:41 babijoee >.<"
04:42 stinebd do you need new ones?
04:42 stinebd i don't think much has changed for diam/raph
04:43 stinebd looks like nothing, actually
04:45 babijoee i wanted to try mr.pippy ram hack
04:45 babijoee but he reverted it :(
04:45 stinebd he reverted it
04:45 stinebd yeah, it was unstable
04:45 stinebd i couldn't get too far after boot with it
04:46 babijoee ah ok
04:46 babijoee did you see alot more ram?
04:46 stinebd didn't even check actually
04:46 babijoee lol
04:46 stinebd the freezes let you know it was working ;)
04:49 babijoee hahah ok
04:55 babijoee shadowline you on here?
04:55 stinebd i've been testing nohz today to fair results
04:55 babijoee nohz is idle sleep?
04:55 stinebd nope, but it enables more idle sleeping
04:55 babijoee awesome
04:55 stinebd it's a tickless system. timer interrupts don't fire unless the kernel needs them to
04:55 babijoee so whats can you tell so far
04:56 babijoee oh ok
04:56 stinebd chamonix had turned it off in kernels before because the clock went awry
04:56 we350z hello friends
04:56 babijoee awry?
04:56 stinebd it would stop running and end up slow by minutes after a few hours
04:56 babijoee hi we350z
04:57 stinebd so after say 8 hours it would be 5 minutes behind actual time
04:57 babijoee oh ok
04:57 stinebd i haven't had such an issue, which i'm hoping is a result of the system using idle sleep now
04:57 babijoee traded for extra battery life i think people wouldnt mind
04:57 stinebd however, i have had an issue with usb
04:57 babijoee yeah?
04:58 stinebd usb debugging (and usbnet) don't work after the first idle sleep
04:58 babijoee hmmm =/
04:58 we350z so i did finally get one of the older kernels on UTB to partially boot on my device using mtype 1805 today
04:58 stinebd but charging still works
04:58 stinebd it may be vbus detection problems
04:58 babijoee probally
04:58 stinebd but that's kind of a stopper for me
04:59 stinebd since it essentially breaks wired tethering
05:00 we350z and MrPippy enlightened me to some things including the mmu dumps from the haret console, and modifying the current htc-msm-2.6.27 source to support this device
05:01 we350z basically the panel and sd are not being initialized for mtype 2426, and I didn't know that enabling the early fb log was an option in the defconifg
05:03 we350z but he is saying memory management does not look like an issue based on the dumps
05:03 we350z so looks promising
05:05 babijoee shadowline was able to get livewallpapers working :) i wonder how.. damn hes not replying to my pm
05:07 we350z livewallpaper is cool
05:09 polyrhythmic babijoee: you're not using version 4 of squashfstools, are you?
05:09 stinebd squash4 won't work with 2.6.27
05:10 polyrhythmic exactly
05:17 we350z is there a way for me to tell what source this kernel was compiled from?
05:18 we350z or any kernel for that matter?
05:18 stinebd boot it up and check uname -r i guess
05:19 stinebd if they put commit id's in the kernel version
05:19 stinebd glemsom kernels do
05:20 we350z ther in lies a problem
05:20 we350z it goes into some sort of infinite loop
05:21 we350z but its the only one i've even been able get to output anything to the screen and at least start booting...
05:21 we350z so it would be great to compare the state of the source at that time (if available) to the current source
05:22 we350z if its even linux-on-qualcomm-s-msm/linux-msm:htc-msm-2.6.27
05:22 we350z which i assume it is but I have no way of really knowing for sure
05:23 we350z can the kernels be decompiled?
05:23 parmaster are there maybe symbols you could look for in gdb?
05:24 we350z gdb?
05:24 we350z i wont be able to get to a prompt...
05:24 parmaster the gnu debugger
05:24 we350z oh
05:24 we350z on my pc?
05:24 parmaster if you have linux running
05:25 we350z i have a linux vm - but how would i attache the kernel to the debugger?
05:26 we350z im not sure what your saying
05:27 parmaster
05:27 we350z suppose i could download the source from around that timeframe and compile it
05:27 we350z kind of a shot in the dark
05:27 parmaster just read that
05:28 we350z interesting
05:28 we350z thanks!
05:29 parmaster again, i'm not sure you'll be able to properly identify it anyway
05:30 we350z hmm
05:31 we350z might be easier to seek out the people who are credited with it
05:32 parmaster gdb is awesome though
05:32 we350z that does look very cool
05:32 we350z does it actually load the kernel on the host pc?
05:32 we350z like some sort of vm or something?
05:33 we350z im not sure how it would do that
05:33 we350z arm emulation?
05:35 parmaster it has different stubs
05:35 parmaster You need to get yourself a GDB that is capable of understanding you target architecture. Often this come with you cross-compiler but if you have do compile it yourself you need to understand the difference between --target and --host configure options. GDB will be running on host(read x86) and will be able to understand target (read armv6). with that you might also want to have the gdbserver that can serve as stub for you user land debugging.
05:35 we350z pretty cool
05:35 AstainHellbring hi
05:36 we350z hello
05:36 parmaster 'lo
05:39 parmaster we350z, anyway there's different things to reverse engineer linux kernels/binaries
05:40 parmaster gdb is the most logical if its a kernel tho
05:59 we350z ok
06:00 tmzt it's armulator based
06:00 we350z so i guess this is based on DZO's kernel
06:00 we350z he is on xdad
06:00 tmzt hey par
06:00 we350z i guess he is the "godfather" of HTC android ports from what i read?
06:01 tmzt we350z: what diamond?
06:01 tmzt who??
06:01 we350z dzo
06:01 we350z this guy
06:01 parmaster heya tmzt,
06:01 parmaster how goes
06:01 tmzt there's no way that there's only one kernel that works on your phone
06:01 tmzt ok
06:01 tmzt well, vogue yes
06:02 we350z yeah
06:02 tmzt but android is just linux, so you have to define htc phone
06:02 tmzt or non-A
06:02 we350z maybe i should try some others - i think its chaning the mtype that did it
06:03 tmzt what phone?
06:03 we350z htc treo pro
06:03 tmzt ah
06:03 tmzt so what mtype?
06:03 we350z ill try the latest raph800 kernel again
06:04 we350z 1805
06:04 tmzt which is?
06:05 we350z gsm diamond i think?
06:05 we350z diam100
06:05 we350z i guess raph800 wouldnt support that mtype
06:07 we350z i tried this brand new one
06:07 we350z
06:07 we350z no dice
06:07 we350z gets a white screen
06:09 we350z i'd love to see the code
06:09 we350z it is hard to beleive it would be the only kernel that works
06:09 we350z well partially works
06:10 we350z i suppose this one is worth trying as well
06:10 we350z
06:10 MrPippy we350z: hey
06:11 tmzt how far?
06:11 tmzt what mtype
06:11 we350z whats up dude
06:11 tmzt what date?
06:11 we350z 1805
06:11 we350z 2009.09.09
06:11 tmzt which is for?
06:11 we350z RAPH100 from what i gather
06:11 tmzt it might be the mddi init changes
06:11 tmzt and raph800 or 500?
06:11 we350z mddi ?
06:11 MrPippy that kernel from september 9 is ancient, from ltg:;a=shortlog;h=refs/heads/htc-msm-2.6.27
06:12 tmzt same kernel, different mtype
06:12 tmzt if you can bisect that and see what change broke it it might help
06:12 tmzt but a whole lot has changed since septemeber
06:12 we350z yeah
06:12 we350z i could just keep grabbing and compiling
06:12 we350z see where it breaks
06:13 we350z how do i know this is the same source this dzo guy or whoever compiled it used?
06:13 we350z is this what everyone uses?
06:14 MrPippy thats the git we were using back then
06:14 we350z how would i check that out?
06:15 tmzt bisect is a process
06:15 MrPippy i'm surprised that would boot with 1805 (gsm diam mtype) because it put the framebuffer and other pmem areas in smi2, which the 7501a doesn't have
06:16 MrPippy although you said it didn't get very far
06:16 tmzt it finds the middle and goes down to the change or set of changes
06:16 we350z right
06:16 we350z it mounts the sd card
06:16 we350z then it goes into this infinite loop
06:16 we350z i found a post where someone else had the same problem on another device.
06:17 MrPippy i wouldn't expect mtype 1805 to be able to mount an sd card on your device, the diamond uses internal storage not an sd slot
06:17 tmzt 1805 is diam100?
06:17 we350z i **think
06:17 tmzt yeah it should and there's no detect on diam
06:17 MrPippy yeah
06:18 tmzt so it should just mount if the mmc host is right
06:18 tmzt so on tpgsm it's the same as diam?
06:18 we350z
06:18 tmzt what about the wifi?
06:18 we350z look at #4928
06:18 we350z thats the problem i have
06:18 we350z acutally MrPippy is on that post :-)
06:18 we350z *thread
06:20 MrPippy for diam500 (with internal storage) the gsm mmc pdata wouldn't work, if treopro is similar to raph cdma it should need the cdma mmc pdata
06:20 we350z im just wondering what it is that makes this one even output anything - since no other kernel i have tried so far seems to do so
06:21 we350z tho i havent tried mtype 1805 on that many kernels
06:21 tmzt output what?
06:21 tmzt probably mddi is not inited and that's why
06:21 tmzt or part of it
06:22 we350z wel i get the little penguin picture in upper left and acutally get text output and see that the kernel is trying to boot
06:22 tmzt if you could bisect from that date in ltg/linux-msm to current linux-msm would be very helpful
06:22 tmzt to what point?
06:22 MrPippy maybe it uses gsm diamond-style panel
06:22 tmzt tried rootdelay=5 ?
06:22 MrPippy not sure if diamond-panel even existed back then
06:22 we350z does that slow the message output?
06:23 we350z it flies by quick but from what i could see it hits like found sd card trying to mount then it said something about doing a fsck
06:23 we350z then it hits that loop as described in the xda post i just linked
06:23 tmzt no it lets it detect sd before trying to mount
06:23 tmzt if you get panic: failure to mount root on whatever that's good
06:24 we350z let me boot it again
06:24 we350z its ahrd to see once it hits that loop
06:24 we350z what the last legit message was
06:25 tmzt loop of?
06:25 MrPippy yeah 1805 used raphael-panel in september, thats probably the right one to use
06:25 we350z msm_gpio_enter_sleep gpio
06:25 we350z in that link i posted he explains the loop
06:25 we350z it actually happens pretty quick in the boot process
06:26 we350z how can i slow this thing down? Maybe i can use my digital camera to record it and playback slow
06:35 joee_ live wallpaper is so awesome
06:35 joee_ :)
06:40 we350z ok i took a video of it
07:34 joee_ new build for anyone who would like to test
07:34 AstainHellbring joee_ huh?
07:34 joee_
07:40 TheMuffinMan Downloading the new build now
07:41 we350z
07:47 we350z +mddi_client_nt35399_init
07:53 we350z what does that mean?
07:53 we350z mmc1: new high speed SDHC
07:54 we350z mmcblk0: mmc1:0518 SD16G
07:54 stinebd that would be the sd card
07:55 Tiddo netripper pong ..
07:55 we350z so its detected
07:55 stinebd 16gb
07:55 we350z yep
07:55 stinebd my 1gb card weeps
07:56 we350z :-)
07:56 we350z running an fsck on the SD card
07:56 we350z Failed to mount /system
07:58 we350z why?
08:14 babijoee we350z: is the card format to fat fs?
08:15 we350z fat32 i beleive
08:15 we350z let me check the vid again
08:17 we350z yeah FAT32
08:17 we350z seems to be reporting the correct number of files
08:18 we350z what was someone saying about mddi before?
08:18 we350z is saw this near the very beginning of the messages
08:18 we350z +mddi_client_nt35399_init
08:19 we350z
08:25 parmaster is there an htc codename for the treo pro?
08:26 we350z i have heard a few
08:26 we350z i think they have diff ones for the gsm and cdma version
08:28 we350z panther for cdma, cheetah for gsm
08:28 we350z Palm Treo Pro (Monk) 850e
08:29 we350z
08:29 we350z
08:30 tehtrk_home can I pass kernel parameters like rootflags=data=journal in the default.txt for haret?
08:31 tehtrk_home or more generally, can I pass kernel parameters as if it was a grub kernel line in the default.txt
08:32 we350z i have a lot of tinkering to do
08:34 we350z isnt that what cmdline is for?
08:42 tehtrk_home no, I'm trying to pass in kernel parameters at the boot loader, like in linux if I wanted to make sure acpi was disabled, I could use acpi=off as a boot parameter
08:44 tehtrk_home trying to use journaled data mode for ext3 filesystems on my kaiser
08:45 tehtrk_home android is my new gentoo :-D
08:58 we350z h
09:09 tehtrk_home didnt work, installer ran mke2fs again
09:11 tehtrk_home oh well, I'll get there
09:18 we350z i need to do some hackery
09:54 Markinus phh: Hi, I removed all CONFIG cases and replaced with a function call, all is checked in . .
11:06 phh [01:04:36] <tmzt> what's EBIN? <---- base of bank 2 EBI1
11:08 phh [01:14:35] <makkonen> hmm. getting a call while eclair is booting is not a great experience. vibrate seems to be stuck on now. and the device is hung. <------ start wih airplane on ?
11:08 phh in wimo
11:08 phh that's what I do, it prevents many crashes.
11:08 phh [01:16:45] <ToAsTcfh> phh: so are u guys using msm_camera or did u fix up libcamera to work with the other? <------ it's still not working but it will be msm_camera, sorry :p
11:09 phh [01:21:54] <emwe> MrPippy: i got 64MB + 51MB working. but that was pure luck it seems. isn't there said to be some radio basband stuff somewhere in the banks whcih sould not be "touched"? <----- only SMI and bank1, which are already taken care of
11:10 phh [01:23:14] <MrPippy> but we've always been able to use the 107MB in bank1, not sure if this will help you use more of your bank1 <----- unless you remove gpu0 (why not hum.) you'll stay with 107
11:11 phh [01:25:46] <emwe> oh ah sorry... mixed that up... GPU1 can be reloacted to "EBIN". that worked. but you couldn't claim it's "ususal" place in EBI_BASE+107 with bank0 <------ really ? I think someone already tried before mmm
11:12 phh " Revert "Backport "ARM: Fix pfn_valid() for sparse memory" to fix bank2"" baaaaah
11:12 phh [01:53:56] <MrPippy> aww it finally crashed after an hour...huge improvement though <------- because of the memory ?
11:14 phh [04:37:18] <babijoee> -sudo genext2fs -d ./system -b 100000 -a system.ext2 <---- Oh i didn't knew that, cool :)
11:16 phh Markinus: uh, where did you chcked it in ?
11:16 phh I can't see any commit on gitorious /o\
11:16 phh oh well glemsom has seen them.
11:16 Markinus yes, I don'T to, but if oyu arte going to the sourcetree in bitorios the files are modified . .
11:17 phh ok
11:17 Markinus maybe a error in bitorious . .can you pull it?
11:17 phh glemsom can so I think it's their cron refresher which is down
11:20 Markinus i didn't changed the raphael-rfkill, I think we replace this in the future with gps-rfkill on all devices . .
11:20 phh no
11:20 Markinus gps
11:20 Markinus sorry
11:20 phh raphael-rfkill is for bluetooth
11:20 Markinus raphael-gps
11:20 phh oh right.
11:21 phh I'm not sure how we will do that
11:21 Markinus did this works on raphael?
11:21 Markinus gps-rfkill?
11:22 phh bah it starts/kill gps, but we have no way to retrieve datas yet :p
11:22 Markinus hm
11:22 phh well we don't know how we are supposed to retrieve data
11:23 Markinus is there no way to dump it under winmo?
11:24 phh I don't know, I leave RE to cr2 :p
11:24 Markinus ;-)
11:25 phh Markinus: hum, htc_fb_console works with your changes ?
11:26 phh I already tried such a change, without much success
11:26 phh oh right you're still using static sized arrays, might work
11:27 Markinus we have to test in on all devices . . . I hope I found all places . .
11:28 phh well if it works for your device, it will be only minor bugs on other devices
11:28 Markinus yes, here it's working . .
11:29 phh fine then, htc_fb_console is quite useful to debug sometimes :D
11:30 phh Markinus: hum, I don't really like doing a switch() in every single call function
11:31 Markinus it'S only the first time?!?
11:31 phh not the amss_para
11:31 phh handle_adsp_rtos_mtoa_app
11:31 phh in this function
11:32 phh amss_para is in qdsp5 ? uh ?
11:33 Markinus I wasn't shure where I have to put this in, the most parameters are using there, so I put it there. .
11:34 phh even, everything that is in qdsp5 should be used only by qdsp5
11:35 chamonix hey guys
11:35 phh hey chamonix
11:35 chamonix hey phh... still up or up again? :)
11:36 phh lol
11:36 phh up again
11:36 Markinus handle_adsp_rtos_mtoa_app: this was a problem, but a switch case is fast . . .
11:36 chamonix HeHeHe
11:36 phh Markinus: I don't know how often it is called
11:36 leobaillard chamonix: [00:52] <phh> cr2_: really ? sounds weird to me ... anyway, I have to go to sleep.
11:36 phh :)
11:36 chamonix hey to u to leobaillard :P
11:36 leobaillard hey to you too :)
11:36 phh Markinus: anyway your patch seems great
11:37 Markinus good :-)
11:38 chamonix does anyone know if there is a way to have an apk "expose" it's API, means the Intents it would serve?
11:39 phh chamonix: you'd better ask on #android :p
11:40 chamonix w00t? y mean there are no Java ppl here?
11:40 phh I don't think so :p
11:40 chamonix pfffff
11:40 chamonix :)
11:50 phh stinebd: so what about nohz ?
12:14 phh I love qualcomm.
12:14 phh their 3D userland driver maps the 0x0 address.
12:14 Markinus ohh
12:15 phh (and yes our kernel prevent the userland from doing such a stupid thing)
12:15 Markinus :-)
12:16 phh it still doesn't work though.
12:19 phh I still get "[ 1706.920959] Unhandled fault: external abort on linefetch (0x806) at 0x00000000" ...
12:21 Markinus difficoult to find such error without sources ..
12:22 Markinus close source driver, or?
12:22 phh yes
12:35 emwe hiho
12:39 dcordes_ emwe, hi
12:42 emwe hi dcordes
12:42 stefan__ hi all, do you guys already know
12:42 stinebd phh: still going very well for me
12:42 stefan__ maybe it could help to get camera working
12:42 stinebd i have to work all weekend so all i can do is keep trying it like this
12:49 emwe phh: is gitorious not up on par with displaying the most recent activities?
12:49 emwe MrPippis push last night wasn't shown as well...
13:10 phh stefand1: not at all.
13:10 dcordes_ phh, still didn't get linuxtogo to make friends with you?
13:10 phh I haven't tried for a while
13:11 dcordes_ maybe the problems are gone now
13:12 phh yes maybe. but gitorious has a nice feature that ltg doesn't have, and that is widely used now. the ability to clone and host it a repo
13:15 dcordes_ phh, I put a link in the wiki. might still be useful to add everything in ltg
13:19 dcordes_ phh, sorry what do you mean, ability to clone and host it a repo?
13:20 phh see personal clones in
13:22 dcordes_ you can do local branches in git
13:23 dcordes_ I don't see the need for 25 different 'private' branches
13:23 phh dcordes_: at least two of these repos are used by multiple persons at the same time ...
13:28 phh hum it seems I can push this time.
13:28 phh let's see if it can do the whole push.
13:28 phh again frozen after "Total 1460 (delta 1128), reused 319 (delta 258)"....
13:29 dcordes_ phh, what is bma150 ?
13:30 phh g-sensor for topa (and rhod I think
13:30 phh )
13:30 dcordes_ ok
13:33 dcordes_ bosch ?
13:33 phh yes
13:34 phh but not like htc hero
13:34 phh htc hero has direct SPI link to bma150
13:34 phh topa/rhod are like N1, the SPI link is over microp
13:34 dcordes_ can you document it in the wiki?
13:35 phh I don't know how it works, I just copy pasted N1's driver ...
13:35 phh (and it's not working yet anyway)
13:35 emwe phh: is there android-side support needed in order to get the bma150 going?
13:35 phh emwe: first linux-side :p
13:36 phh if the kernel driver is "GNU/Linuxish", nothing will have to be done since I've already done one for kxsd9
13:37 emwe but android has to ioctl() the driver somehow, right?
13:38 phh depends on what the driver is
13:38 emwe is it watching some dev-fs entries?
13:38 phh just read() on /dev/input/eventX
13:39 phh ok we need some writes to /sys/blabla to enable, set rate and scale
13:39 emwe ah, the kernel side stores it's activities in /dev/input/eventX and android-side is polling, right?
13:40 emwe so your recent port of th bma150 over spi actually has to push it's events to /dev/input/eventxyz, understood correctly?
13:40 emwe if it is not actually doing so, yet..
13:40 phh emwe: yes for the first sentence
13:40 phh no for the second one
13:40 phh well
13:40 phh depends on what you mean with "has to"
13:41 phh that's what it will eventually do, but correctly it doesn't.
13:41 emwe phh, ah ok, i see. thanks for the clarification.
13:41 phh s/correctly/currently/
13:41 apt phh meant: that's what it will eventually do, but currently it doesn't.
13:43 cr2 phh: any progress with the cam ?
13:43 phh no ... I still get an I2C error ...
13:43 cr2 phh: can you paste the diff ?
13:44 phh the diff of what I did yesterday ?
13:44 cr2 yes
13:44 phh
13:45 cr2 thanks
13:47 phh I'll try tracing libgps on my phone.... maybe with some big hacks it might do something correctly
13:48 cr2 can we add the vreg list to msm_camera_sensor_info somehow ?
13:49 phh bah
13:50 phh yes we can, but I'm not really fond of that
13:51 cr2 why ?
13:51 phh no particular reason
13:52 cr2 then you will have if_machine_foobar() switch
13:52 phh yes yes I know the problem with not using that ...
13:52 cr2 which was already plagueing the arm kernels long time ago
13:53 phh I'd better make an init cam callback.
13:54 cr2 i can document the full cam init, you then decide
13:54 cr2 and deinit
13:55 cr2 i'll also try to dump the gpio states now
13:56 phh hum I guess I'll just rename camera_gpio_on
13:56 phh no it's not.
13:56 cr2 do you have a link to the mi380 driver ?
13:56 phh bah.
13:57 phh mt9d112 ?
13:58 cr2 yes
13:58 cr2 raph mi380 code looks very comprehensive
13:58 phh
13:59 phh but I don't think it's even near this device actually
13:59 phh no android device has mddi camera
13:59 cr2 phh: mddi camera ?
13:59 phh vga camera is on mddi bus no ?
14:00 cr2 the cam gpio setup is the same for both
14:00 phh uh ? mddi client uses the same gpios ?
14:00 cr2 something to be clarified
14:00 cr2 both cams use the same gpios ?
14:01 cr2 you mean the mddi mdc interface ?
14:01 phh yes
14:01 cr2 needs to be traced. i have no idea.
14:01 phh ok
14:04 phh I can confirm that codeaurora's driver uses exclusivly parallel camera, no mddi camera
14:05 phh (switching from parallel to mddi is just a bit, and they always set this bit to 1.)
14:06 phh anyway if you say the I2C controls are easy, we might be able to do a driver.
14:07 phh anyway. vreg stuff.
14:07 cr2 i'll document it.
14:12 phh cr2: in mt9t013 driver, there is a "mt9t013_client->addr = mt9t013_client->addr >> 1;", so I'm just thinking ... maybe I should use the address without >>1 ?
14:13 cr2 yeah. looks silly.
14:13 cr2 what is g1 doing ?
14:13 phh #if defined(CONFIG_MSM_CAMERA) && defined(CONFIG_MT9T013)
14:13 phh {
14:13 phh I2C_BOARD_INFO("mt9t013", 0x6C),
14:13 phh },
14:13 phh #endif
14:13 phh #ifdef CONFIG_SENSORS_MT9T013
14:13 phh {
14:13 phh I2C_BOARD_INFO("mt9t013", 0x6C >> 1),
14:13 phh },
14:13 phh #endif
14:14 phh <3
14:14 cr2 this is dumb because all other drivers use real id
14:14 cr2 haha
14:14 phh and MSM_CAMERA is for V4L driver, the one I use
14:14 phh so I have to use 0x6c
14:15 cr2 and you use the driver from I2C_BOARD_INFO("mt9t013", 0x6C), tree ?
14:15 phh yes
14:15 cr2 i'd say, fix the driver
14:15 phh right.
14:15 cr2 and use id=0x36
14:16 cr2 as it should be.
14:16 cr2 using id*2 is a convenience for mmutrace and wince ioctl RE
14:17 phh maybe the RE their own devices :D
14:17 cr2 LOL
14:19 phh cr2:
14:20 phh hum, s/camera_set_state/set_state/
14:20 cr2 595
14:20 cr2 /* Input MCLK = 24MHz */
14:20 cr2 596
14:20 cr2 msm_camio_clk_rate_set(24000000);
14:20 phh anything wrong with that ?
14:21 cr2 is it the same as enabling vfe and vfe_mdc clocks ?
14:21 phh yes
14:21 cr2 ok
14:22 phh oh wait no
14:22 phh it just sets their clock rate
14:22 cr2 the driver looks like from wince
14:23 cr2 in the structure
14:23 cr2 that's good :)
14:23 phh that's why there are some i2c_txdata things then
14:24 phh (which hurts me a lot.)
14:24 cr2 ok
14:25 phh cr2: does the command IDs looks like the one you gathered ?
14:26 phh [ 15.994628] msm_i2c msm_i2c.0: Error during data xfer (-5)
14:26 phh [ 16.025817] mt9t013_i2c_txdata failed
14:26 phh [ 16.042144] i2c_write_w failed, addr = 0x301a, val = 0x1009!
14:26 phh failed again.
14:27 phh [ 15.978912] msm_i2c msm_i2c.0: error, status 43c8
14:27 phh let's see what this status means.
14:28 cr2 you gpio setup may be incomplete
14:28 cr2 the need to mux properly between mi380 and mt9p013
14:28 phh this i2c status means I2C_STATUS_PACKET_NACKED&I2C_STATUS_FAILED&I2C_STATUS_BUS_ACTIVE&I2C_STATUS_BUS_MASTER.... nothing new in it then.
14:29 phh cr2: I'm pretty sure mi380 doesn't use this gpio
14:29 phh still the gpu mux might be wrong.
14:30 phh currently it sets gpio from 2 to b to 5, same for d and e. gpio c and fgets 1c5
14:30 phh s/fgets/f gets/
14:30 apt phh meant: currently it sets gpio from 2 to b to 5, same for d and e. gpio c and f gets 1c5
14:30 cr2 that's why i'd like to dump the gpio states
14:31 cr2 the dll depends on some runtime smem setting, so it's easier just to check the real thing
14:31 cr2 note to myself. i also need to find the HCI buffer location, so we can mmutrace BT.
14:32 cr2 omg
14:32 cr2 found old kaiser dumps :)
14:32 phh (I pushed the use of set_state)
14:35 cr2 phh: btw, 4 crc32 versions in wikipedia
14:36 phh hci buffer location is for sound ?
14:38 cr2 fmradio
14:38 cr2 sound is routed as raw pcm. i think
14:38 cr2 bt sound
14:39 phh just an hci command to tell him to transfer pcm sound then ?
14:42 phh cr2: hum, in wiki there is a 0x20 (msme2à pmic enabled for camera, but you didn't report it yesterday, could that be the problem ?
14:42 cr2 you can't mmutrace the TX and RX registers, but they are buffered in ram
14:42 cr2 phh: for which device ?
14:42 phh raph100
14:42 phh
14:43 cr2 (8000),100,400000 and 20
14:43 cr2 wlan(on rhod for VCM), gp1, gp5 and gp3
14:43 cr2 the wiki isfor raph800 ?
14:43 cr2 weird device :)
14:44 phh uh ? 0x20 is gp3 ?
14:44 phh oh I miscalculated my bits.
14:44 cr2
14:45 cr2 "gp3",RUIM2 5 0x20
14:46 cr2 phh: MT9V113 aka MI380
14:47 cr2 phh: not 112, so it needs to be checked.
14:47 phh cr2: anyway vfe code is wrong
14:48 cr2 not everything at once :)
14:48 phh I mean camera vfa interface
14:49 cr2 we need some special case for IMEM and VFE_MDC clk
14:51 phh I added wlan vreg, still same problem.
14:51 cr2 wlan vreg is for rhod only
14:52 cr2 i think your gpio mux is wrong.
14:52 cr2 need to buy some food. bbl
16:09 cr2 phh: connected raph100
16:09 phh cr2: I'm trying to use android's libgps on my diam and trace what it does.
16:10 cr2 ok
16:11 cr2 from raph100
16:12 cr2 watch gpios 0
16:12 cr2 err, wrhong window
16:12 phh :)
16:13 cr2 hm. haretconsole may print linear pin numbers instead
16:14 cr2 hardcopy :)
16:20 cr2 1e -> 1-14
16:21 cr2 6d -> 5-2
16:21 cr2 in idle
16:21 cr2 out1-14=1
16:22 cr2 weird
16:23 cr2 only 5 banks in haretconsole ?
16:23 cr2 seems 7200 biased
16:27 cr2 +0x844 in5
16:27 cr2 +0x850 out5
16:27 cr2 says wiki
16:30 leobaillard OMFG, I found a most disturbing fact about RMS :s I thought he was weird and so on, but not up to that point ! "For personal reasons, he does not actively browse the web from his computer; rather, he uses wget and reads the fetched pages from his e-mail mailbox." (wikipedia)
16:31 tmzt we're missing gpio traces this whole time?
16:31 tmzt wget coesn't use cookies, curl does/can
16:31 cr2 leobaillard: wget should go over tor
16:32 cr2 tmzt: i don't have the latest haret git, but i think so
16:34 cr2 omg
16:34 cr2 defineMsmGpios
16:36 cr2 intr_en is not traced too
16:38 cr2 addlist gpios p2v(0xa9200854)
16:38 cr2 addlist gpios p2v(0xa9200844)
16:38 cr2 addlist gpios p2v(0xa9200854)
16:38 cr2 hm. very newbie friendly :)
16:46 phh [ 334.524719] rr_write: cannot write to 3000005b:-1875374787 (bound to 3000005b:0)
16:46 phh pffff
16:46 phh if they did that correctly, my simple 'ln -s' would have been enough
16:47 cr2 %d for -1875374787 ?
16:47 phh should be %x I guess
16:47 cr2 you also need to check the PROC numbers
16:47 phh but that's not the problem
16:47 cr2 they are different between amss
16:47 phh I know
16:47 phh but first I need him to at least try sending commands :p
16:47 cr2 topa is different from raph
16:47 cr2 ok
16:53 phh -_-'
16:53 phh I crashed ARM9.
16:54 tmzt phh: did what correctly?
16:54 phh tmzt: forcing the send of the rpc message even if the version doesn't match
16:54 phh so that was quite expectable.
16:55 tmzt no, the ln -s
16:55 phh oh
16:56 phh the rpc id/rpc ver shouldn't be sent by the userland by any other way than the device it opens
16:56 phh imho
16:56 Markinus phh: a chacked in a small patch for my patch, the starting of adsp driver was wrong, there is a constant string needet . .
16:56 Markinus checked
16:57 phh [ 226.293090] [RPC] CALL to 3000005b:0 @ 0:00000001 (44 bytes)
16:57 phh cr2: means funtion =1 ?
16:58 cr2 phh: 0:1 is some main arm9 server
16:58 tmzt yes, that would make sense but I guess the rpcrouter is a little too specific
16:58 phh ah.
16:58 tmzt probably so wire can change with version
16:58 cr2 phh: binpatch ;)
16:58 tmzt cr2: the rpcr istself?
16:58 phh cr2: argh.
16:58 cr2 there is no free gps software for g1 ?
16:59 tmzt can't we get these things dumped in a pcap like format
16:59 tmzt don't think so
16:59 tmzt I think swetland might be interested in a nmea driver like yours, amss is removing nmea in later versions
16:59 cr2 phh: use ida
16:59 tmzt nmea kernel driver I mean
16:59 phh :'(
17:00 cr2 tmzt: nmea kernel is dumb
17:00 cr2 tmzt: /dev/smd27 should be enough
17:00 tmzt hmm, so userspace should use pdapi?
17:00 tmzt yes, but I think 8k's don't have smd27
17:00 cr2 tmzt: leo has it
17:00 tmzt fully available?
17:01 tmzt maybe it's 86xx then
17:01 tmzt like other cdma versions
17:01 cr2 tmzt: i've even learned NetRipper position from his smem dump. never post your smem online :)
17:01 tmzt icbm
17:01 phh lol
17:01 cr2 yeah, these dirty linux hippies
17:02 tmzt I mean it doesn't have to be started/stopped over rpc?
17:02 phh tmzt: you can do a kernel-land rfkill
17:02 cr2 tmzt: we don't know yet
17:05 cr2 phh: gpio 1-14 aka 0x1e aka 30 is vga/main mux
17:05 phh mmmmmmmmmmmmuh ?
17:05 cr2 idle=1, vga(mi380)=1, main(m9t013)=0
17:06 cr2 5c is main reset, 5d is vga reset
17:06 cr2 6d is some smem-dependent magic. my value=0, but it may be different elsewhere
17:06 cr2 6b=0
17:07 cr2 5b is VCMPDP
17:07 cr2 =1 in all my tests
17:09 cr2 0xf=mclk should be 0x1c4 with clock enabled
17:09 tmzt vga is mddi?
17:10 tmzt I thought it was cfa or something
17:10 cr2 tmzt: uses the same CAM bank0 gpio pins.
17:10 phh tmzt: it's not mddi then.
17:10 tmzt how many pins?
17:10 cr2 11-2+1
17:11 cr2
17:11 phh Captnoord: how do you fix fonts in IDA ?
17:12 cr2 tmzt: mi380 is 113 , and there is 112 driver
17:12 cr2 should be doable.
17:15 cr2 omg, now i get the whole ulpi traffic
17:18 dcordes-kais NetRipper ping
17:18 phh dcordes-kais: I tried pushing, still not luck
17:19 cr2 dcordes-kais: ltg for clean drivers
17:19 dcordes-kais yea
17:19 dcordes-kais but it's bad that phh can't acces
17:19 cr2 we need to cooperate with NetRipper on sane dex driver with softirq support.
17:20 dcordes-kais I will also get a leo. just tried it out in a vf shop
17:20 cr2 lol
17:20 tmzt how is that? for 8k only?
17:20 cr2 dcordes-kais: why not nexus ?
17:20 tmzt like it better than nexus?
17:21 dcordes-kais is that also the mrketing name?
17:21 cr2 tmzt: dex driver is the same. on 8k smem is at a different place, and dex offset too
17:21 tmzt Nexus is an 8k device from HTC sold by Google
17:22 tmzt it's android based with a proper kernel
17:22 tmzt serving as basis for hd2 and lgexpo support
17:22 cr2 with smaller amoled lcd, and biiger sdram
17:22 tmzt also has OLED screen and capacatative input
17:22 tmzt 512!!!
17:22 cr2 tmzt: no second cam
17:22 tmzt or no?
17:22 tmzt that's Pre Plus
17:23 cr2 tmzt: tvout ? fmradio ?
17:23 tmzt probably fm but no software support
17:23 tmzt could be eu reg reasons
17:23 cr2 does androed support second cam, tvout and fmradio ?
17:23 tmzt if it has 3.5 there's no reason to believe it's not routed for fmant
17:23 cr2 tmzt: on bcm4329 ?
17:24 cr2 one more reason to trace BT
17:24 phh cr2: tatoo has fmradio support
17:24 tmzt I don't know for sure, we have to try it
17:24 phh but no source code, and involve framework changes.
17:24 cr2 phh: what is tatoo ?
17:24 phh and proprietary bluetooth stack
17:24 phh cr2: a low-end htc android phone
17:24 dcordes-kais cr2 htc androed devce
17:25 cr2 and no source ?
17:25 cr2 do we have hw specs for it ?
17:25 tmzt it's a magic
17:25 tmzt I think
17:25 tmzt just the smaller ram version
17:26 phh uh ?
17:26 dcordes-kais is nexus available already?
17:26 phh dcordes-kais: yes
17:26 cr2 tmzt: so it's ti brf63xx ?
17:26 dcordes-kais google sells it online?
17:27 phh cr2: yes
17:27 phh dcordes-kais: yes
17:27 tmzt cr2: which? tatoo? I think so
17:27 cr2 phh: then we may patch the uart driver, and dump all fmradio stuff
17:27 cr2 from this "proprietary" bt stack
17:28 phh cr2: first we have to get this stack working ...
17:28 tmzt npelly should have patches for that
17:28 tmzt dumping uartdm
17:28 cr2 phh: where can i get this stack ?
17:28 phh I don't know
17:28 tmzt what stack?
17:28 cr2 tmzt: taoo rom
17:28 phh tmzt: btips
17:29 phh cr2: and but I don't know if it needs other libs
17:30 tmzt is this yours? what phones use it?
17:30 tmzt how does android control it?
17:30 phh I think it uses dbus
17:30 cr2 phh: wget :)
17:31 phh tmzt: anyway you have the daemon and the cli, if it doesn't lack any lib it's enough :p
17:31 phh I understand just nothing to IDA :(
17:31 dcordes-kais nexus doesn't seem to available only in uk
17:32 dcordes-kais lol rright
17:33 dcordes-kais > library
17:33 tmzt uwi 6
17:35 cr2 phh: done with cam gpio logs
17:36 cr2
17:36 cr2
17:36 cr2
17:36 cr2
17:36 cr2
17:36 phh ok so you need nothing else
17:36 tmzt those use dbus? is it emulating bluez?
17:36 tmzt strings should be eboungh to see what paths it uses
17:37 cr2
17:37 tmzt also, is this tatoo only?
17:37 tmzt hmm, so they don't use audioflinger for routing?
17:37 phh tmzt: no, it's the same stack for almost everyone
17:37 cr2 tmzt: should work for brf63xx
17:37 tmzt all android?
17:37 cr2 tmzt: you have another chip :)
17:37 phh tmzt: they do use audioflinger for A2DP routing
17:37 tmzt I know :)
17:37 tmzt but not sco?
17:37 phh no
17:38 cr2 omg
17:38 cr2 Cmd_FmTX_SetECC
17:38 phh cr2: how do I change a value in IDA ?
17:38 cr2 phh: which value ?
17:39 phh "version" value here,
17:40 phh it's directly in the code
17:40 tmzt hex editor?
17:40 tmzt .db ?
17:40 cr2 find the offset
17:40 cr2 and use 'mc'
17:40 phh ah. ok.
17:40 phh I thaught I could do it though IDA directly
17:41 cr2 f3,f4 afair
17:42 cr2 GetRdsAfSwitchMode
17:42 cr2 SetRdsAfSwitchMode
17:42 cr2 SetRdsGroupMask
17:42 cr2 GetRdsSystem
17:42 cr2 SetRdsSyste
17:42 cr2 no tmc ?
17:43 tmzt rds is just an osi layer for tmc?
17:43 tmzt a transport or so
17:43 cr2 hm. yes
17:43 cr2 found it
17:43 cr2 GetRdsTrafficCodes
17:43 cr2 SetRdsTrafficCodes
17:44 tmzt that's a mask?
17:44 cr2 tmzt:
17:45 cr2 phh: are there any .bts (or bin) files in the rom ?
17:45 phh oh yes
17:45 cr2 pm
17:45 phh it needs pm ?
17:46 cr2 i'd liketo see them
17:46 cr2 BTL_SetModuleDebugLeve
17:47 phh{fm_rx_init_1273.2.bts,fmc_init_1273.2.bts}
17:47 dcordes-bib_ hmmm
17:47 cr2 Command is not an H4 command packet!
17:47 phh there is also{tiinit_5.3.53.bts,avpr.bts}
17:48 cr2 .bts dump works
17:48 cr2 # Compatibility : BL6450 2.1, BL6450 3.0, WL127x 2.0, NL55xx 2.0 ROM
17:49 cr2 nice bts comments there :)
17:49 tmzt wl :)
17:50 tmzt wifi/bluetooth are in same chip on g1,etc.?
17:50 cr2 #release Cortex Reset
17:50 cr2 HCITOOL: hcitool cmd 0x3f 0x301 0x00 0xc0 0x1b 0x00 0x08 0x00
17:50 cr2 tmzt: not on rhod/leo
17:50 tmzt what is that?
17:50 cr2 tmzt: the rest are the same
17:50 tmzt rhod is bcm
17:51 tmzt wl127x has a ti bt ip core?
17:51 cr2 rhod=bcm4325,leo/nexus=bcm4329
17:51 tmzt or it is ti
17:51 cr2 no idea
17:51 tmzt 4329 adds fm transmit or just non-mimo N?
17:52 cr2 phh: looks like it's for 6450 chip
17:52 phh not 6350 ? baah
17:52 cr2 phh: but we have the reference data in wiki
17:53 cr2 i doubt they have changed hci-2-i2c api
17:53 tmzt what is Cortex Reset?
17:53 tmzt unless the new core is cortex-m
17:53 cr2 tmzt: .bts comment
17:53 tmzt yes but it makes no sense
17:54 tmzt options, it's cortex-m or arm9 is supposed to send that meaning wifi/bt can suspend apps processor
17:54 tmzt unlikely
17:54 cr2 # Cortex Reset (hold in reset while loading IMEM)
17:54 cr2 HCITOOL: hcitool cmd 0x3f 0x301 0x00 0xc0 0x1b 0x00 0x09 0x00
17:54 cr2 tmzt: bt/wifi has its own arm processor
17:55 phh in arm eabi, the return value is in R5 ?
17:55 cr2 #turn on COREX clock (should read 0xnnn9)
17:55 cr2 phh: r0
17:56 phh BL j_clnt_create
17:56 phh MOVS R5, R0
17:56 phh CMP R0, #0
17:56 cr2 hm. eabi. sorry
17:56 tmzt cor execute?
17:56 phh ah.
17:56 cr2 #turn on COREX clock (should read 0xnnn9)
17:56 cr2 HCITOOL: hcitool cmd 0x3f 0x301 0x10 0x60 0x1a 0x00 0x09 0x00
17:58 cr2 #set IO17 to be the BT RF-Activity pin
17:58 cr2 HCITOOL: hcitool cmd 0x3f 0x11d 0x04 0x01 0xe9 0x05 0x00 0x00 0x00 0x00 0xff 0xff 0xff 0xff 0xff 0xff 0xff 0xff 0xff 0xff 0x05 0x01 0x00 0x01 0x02 0x01 0x00 0xff 0x00
18:01 cr2 i think we have these already
18:01 tmzt if linux ever gets btphy support this could be helpful
18:05 cr2 and XTRA_* things are not in our code
18:05 phh I know android is stupid, but why do they call clnt_create() twice with the same params uh ?
18:05 cr2 PDSM_SET_PARAMETERS seems to be important
18:05 cr2 phh: RAD
18:05 phh rad ?
18:05 tmzt XTRA is the gps calibration/ephemeris thing?
18:06 cr2 RapidApplicationDevelopment ;)
18:06 phh ok
18:06 phh cr2: you have btw ?
18:06 tmzt uh, probably because an amss client version did it
18:06 phh if you want to have some fun with it.
18:06 cr2 tmzt: it's something qualcomm specific
18:06 tmzt xtra I want
18:06 cr2 phh: from cliq
18:06 tmzt really want
18:06 tmzt made my touchpro2 gps work
18:07 tmzt if it expires it's useless
18:07 cr2 tmzt: it's some spyware ?
18:07 tmzt but we can do it over DM
18:07 tmzt no, it's ephemeris
18:07 cr2 when you must connect to some qcom server in inet ?
18:07 tmzt the QuicKGPS thing
18:07 cr2 yes.
18:07 tmzt yes, to pull a bin file
18:07 phh cr2: it doesn't have to be qcom
18:07 phh it can be google :D
18:07 tmzt it doesn't send your location
18:07 cr2 it's compressed and encrypted ephemeris
18:08 tmzt (except ip)
18:08 cr2 tmzt: LAC ?
18:08 tmzt it's also not realtime, you only have to download once a week
18:08 tmzt not over wifi
18:08 tmzt I'll try my next update over wifi if I can
18:08 phh tmzt: it does send LAC when over 3G ?
18:08 cr2 works without it for me
18:08 phh ergh
18:08 tmzt I don't know
18:08 cr2 phh: need to trace
18:08 tmzt do we knw the protocol?
18:09 cr2 no
18:09 tmzt it won't send lac if modem is off at least
18:09 phh tmzt: the retrieving of these data are done over http
18:09 cr2 LAC can provide some good inital guess
18:09 tmzt last lac maybe
18:09 cr2 for the search
18:09 tmzt it uses almanac and ephemeris
18:10 tmzt maybe amss gets eph from tower by ip using lac
18:10 tmzt but not alm
18:10 tmzt without qgps gps ttff is slow
18:10 tmzt unusable
18:10 cr2 in the US only
18:10 tmzt yes I think so
18:10 cr2 slower than sirf3
18:10 tmzt but I'm in us
18:10 cr2 lol
18:11 tmzt if you have a local gps receiver you can possibly build a file
18:11 tmzt openmoko was working on it
18:11 cr2 there was sme ?
18:11 dcordes-kais tmzt want to buy a nexus for me?
18:11 tmzt talk to onen about opencellid
18:11 cr2 dcordes-kais: me too
18:11 tmzt dcordes-kais: if I get varafied account this week I think so
18:12 tmzt there are eu countries though that might be better
18:12 tmzt cheaper
18:12 dcordes-kais uk only afaik
18:12 cr2 tmzt: cheaper ? lol.
18:13 tmzt it's not in uk?
18:13 tmzt hmm, I think the pricing is usd everywhere
18:13 tmzt I could be wrong
18:13 tmzt
18:13 tmzt you have vat of course
18:13 dcordes-kais nexus and hd2 prices are similar
18:13 tmzt so it depends on shipping costs
18:14 dcordes-kais tmzt they wont deliver to germany directly
18:14 makkonen hd2 seems like an altogether better device. stupid OSes.
18:14 cr2 tmzt: in the eu you get 2 years warranty
18:14 phh dcordes-kais: then get the phone which can runs both oses ? :p
18:14 tmzt one in us?
18:14 tmzt dcordes-kais: talk to ali?
18:15 dcordes-kais phh NetRipper has the leo already..
18:15 tmzt or ce6 bsp maybe people want to work on that
18:16 cr2 i will prefer nexus. to avoid m$ tax
18:16 cr2 anything else for leo ?
18:17 phh you get google's tax, is that better ?
18:17 cr2 an i will remove all lame geoogle software from it. or course.
18:18 tmzt then just use aosp
18:18 cr2 phh: you think i'll use google apps ?
18:18 dcordes-kais aesop?
18:18 tmzt android open source project
18:18 tmzt build from repo
18:18 phh cr2: no and ... ? google will still have a great profit on it.
18:18 dcordes-kais yuck
18:19 tmzt if they buy it wholesale
18:19 tmzt sure
18:19 cr2 at least they have published some linux source.
18:19 tmzt but I think the conspiracy (or Apple) level things are wrong
18:19 cr2 that must be honored.
18:20 cr2 unlike microsoft with its fcking wince
18:21 dcordes-kais with 1GHz they are slowly moving towards higher usability level
18:21 cr2 LOL
18:22 cr2 dcordes-kais: now they need 4GB ram
18:22 cr2 if msm can support it though
18:23 dcordes-kais board-mahimahi.c
18:23 cr2 phh: have you chacked the FUNC_ numbers ?
18:23 phh uh ?
18:24 phh oh that. currently the problem is about the crashing call to 00000:01
18:25 cr2 awk '{printf("0x%x\n",-1875374787)}'
18:25 cr2 0xffffffff90380d3d
18:25 phh cr2: this is not the problem
18:25 cr2 VERS is 0x90380d3d
18:25 cr2 PROG is 0x3000005b
18:26 cr2 but the FUNCs
18:26 cr2 you need to check these numbers too
18:26 phh the problem is that there is a {0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x30, 0x0, 0x0, 0x5b, 0x90, 0x38, 0x0d, 0x3d, 0x00, ... 0x01} (total = 44bytes) rpc call that kills ARM9
18:26 cr2 phh: which func is crashing
18:27 phh [ 226.293090] [RPC] CALL to 3000005b:0 @ 0:00000001 (44 bytes)
18:27 cr2 phh: can you use 4byte format ?
18:27 dcordes-kais it's a fish so it's a qualcomm evaluation board
18:27 tmzt thought so, but trout is a fish
18:27 cr2 they all are a fish
18:28 tmzt mahimahi I think is nexus but not sure
18:28 cr2 tmzt: it is
18:29 cr2 1 0 2 30005b 90380d3d ...
18:29 cr2 phh: what comes after that
18:29 phh cr2: don't take care of the 90380d3d
18:29 phh I change it to 0
18:29 cr2 it's the FUNC number, then 4x 0x0 and then the payload
18:29 dcordes-kais blub
18:29 phh ok
18:30 phh so func number is 2
18:30 cr2 #if (CONFIG_MSM_AMSS_VERSION == 6120) || (CONFIG_MSM_AMSS_VERSION == 6125)
18:30 cr2 #define CLIENT_INIT 2
18:30 phh I have to map it to 1 then
18:30 cr2 hm. sound reasonable
18:30 cr2 5225 AMSS has 3
18:30 phh ah.
18:30 cr2 it's CLIENT_INIT
18:31 cr2 #elif (CONFIG_MSM_AMSS_VERSION == 5200) || (CONFIG_MSM_AMSS_VERSION == 6150)
18:31 cr2 #define CLIENT_INIT 3
18:31 cr2 the payload should be 0xa
18:31 cr2 and the return value
18:31 cr2 /* TODO: replace with msm_rpc_call_reply() values */
18:31 cr2 #define PD_CLIENT_ID 0xda3
18:32 cr2 hm. not 0xa
18:32 cr2 /* constants */
18:32 cr2 #define INIT_PD 0x2
18:34 cr2 (total = 44bytes) rpc call that kills ARM9
18:34 dcordes-kais cr2 what sort of register is */ack ?
18:34 cr2 /4= 11 values
18:34 cr2 dcordes-kais: where ?
18:35 dcordes-kais cr2 MSM_RPC
18:35 cr2 dcordes-kais: bad style frm me
18:35 cr2 there should be server and client numbers
18:36 cr2 andthe FUNC numbers and VERS numbers depend on the amss version
18:37 dcordes-kais somebody knows which exact kernel snapshot is shipped on the nexus?
18:37 dcordes-kais or how to figure
18:38 tmzt auname -a?
18:38 cr2 dcordes-kais: leo is 555€ ?
18:38 dcordes-kais htc_35mm_jack.c lol
18:38 tmzt and nexus?
18:38 cr2 compared to nexus it sounds expensive
18:38 tmzt yes, they emulate h2w for the internal circuit
18:38 cr2 tmzt: nexus can't be bought here till 1.05.
18:38 tmzt and uk?
18:39 tmzt May?
18:39 cr2 do you know any reasonable site there ?
18:39 cr2 yes. 1 may
18:40 phh cr2: you might offer me one for my birthday then. it's on 11.05 :p
18:40 cr2 phh: raph100 ?
18:40 dcordes-kais cr2 it would suck to give away the 2years of warranty
18:40 phh cr2: ah no. I want you to continue developing on it :p
18:40 cr2 dcordes-kais: used only for the universal.
18:41 cr2 phh: raph100 is precious compared to this new wave. tvout, fmradio and 2 cams.
18:41 phh + enough ram + keyboard
18:42 cr2 yes
18:42 dcordes-kais cr2 I sent in kovsky and wil propably have to do it again
18:42 cr2 but no magnetometer
18:42 cr2 dcordes-kais: it's not "real" htc.
18:42 tmzt front facing will come back
18:43 tmzt when ip gets serious bandwidth
18:43 tmzt or Apple has it :)
18:43 dcordes-kais yea true. i wonder if I can fix kovsky again on my own
18:43 phh cr2: I think I can live without a magnometer.
18:43 phh even if it's a neat feature
18:43 cr2 lol
18:44 phh does TP2 have tvout and 2 cams too ?
18:44 cr2 tmzt: do you have tvout ?
18:44 tmzt I don't know
18:44 tmzt well with cable
18:44 tmzt hold on
18:44 cr2 phh: 2 cams, but the vga cam driverdoes not exist
18:44 dcordes-kais cr2 it will be fairly easy to make a bt kb work on nexus
18:45 phh dcordes-kais: the usability isn't really the same.
18:45 cr2 dcordes-kais: it has usb host ?
18:45 phh cr2: isn't that a native MSM feature ?
18:45 dcordes-kais phh but if you tape it to the device :D
18:45 tmzt yes I have the app on my stock rom
18:46 dcordes-kais cr2 let me check. if not you must write it
18:46 cr2 dcordes-kais: there is ehci-msm driver
18:46 tmzt warning TV out cable (an optional accessory) is needed to output to an external display
18:46 phh rq->vers==0x0;
18:46 phh I hate my keyboard.
18:46 cr2 tmzt: ok. i have this cable
18:47 tmzt what device do you have?
18:47 tmzt it's extusb?
18:47 cr2 raph100
18:47 cr2 yes
18:47 cr2 the pinout is known
18:47 cr2 but you can't have tvout and serial console at the same time
18:47 tmzt I can try a no2chem program and force it on
18:47 phh cr2: most people can live with that :p
18:48 tmzt no, and on Nexus serial is on if vbus is not
18:48 tmzt so probably no tvout
18:48 tmzt maybe without microp firmware update
18:48 cr2 lame
18:48 tmzt but that's a theory
18:48 phh tmzt: maybe the pinout is different ?
18:48 tmzt I think it's the same
18:48 cr2 does leo have tvout ?
18:48 tmzt just microp driving gpio switch
18:49 cr2 must be the same. there are not that much pins to use
18:49 tmzt based on adc of vbus pins
18:50 cr2 leo is such a brick on the photo
18:50 dcordes-kais cr2 it seems like ehci msm is onl in codeaurora
18:51 cr2 dcordes-kais: yes, but it's a standalone driver
18:51 dcordes-kais ah
18:51 dcordes-kais how about trying it on raph?
18:51 cr2 we are too far behind anyway
18:51 dcordes-kais lol
18:51 cr2 gps is more important :)
18:52 dcordes-kais i need to check if i can fix kovs then I can try
18:52 tmzt ehci is?
18:52 phh tmzt: usb2 host
18:52 tmzt not in nexus kernel?
18:53 tmzt yes
18:53 tmzt what do they use on other?
18:53 phh no usb host in nexus kernel
18:53 dcordes-kais nexus has gadget,
18:53 cr2 dcordes-kais: found for 517€
18:54 cr2 nexus/leo are 8250
18:54 tmzt gadget and gadget functions?
18:54 phh tmzt: yes
18:55 tmzt cool
18:55 tmzt no more android function?
18:55 phh bah that's "gadget function" no ?
18:55 phh if not, sorry but it's still the function stuff :p
18:56 tmzt no, the function concept is in gadget now
18:56 tmzt just not the bitmap thing
18:56 cr2 phh: on kais the get_position is a timer callback.
18:56 tmzt so a lot more pdata
18:56 cr2 phh: but i need to check it for raph
18:56 phh ok
18:56 phh I'm having fun with android meanwhile.
18:57 cr2 phh: since your PDSM abi looks like 6125, you may have more rpc functions
18:57 slight can't someone point me in the direction of a channel where i can ask a potentially stupid question about eclari on a raph100? :)
18:57 slight *eclair
18:57 dcordes-kais tmzt what bitmap thing?
18:57 cr2 phh: post the full crahsed pdsm rpc
18:57 phh dcordes-kais: bit 1 stands for adb, bit 2 for mass storage, bit 3 for usbnet, etc
18:58 phh then you write the result in a file and you'll get whatever you asked
18:58 phh cr2: hum I have to revert back to board-htcraphael-gps.c ...
18:59 dcordes-kais bbl
19:00 phh cr2: let me do another try of my kernel hack first and I'll do that.
19:01 cr2 ok
19:01 tmzt dc: the thing that allows enabling multiple functions in a composite device without four pages of structs
19:01 tmzt slight: here
19:01 slight cheers :)
19:02 slight trying to get a > into pterminal so i can dump my dmesg output to a file
19:02 tmzt cr2: xtra protocol might be crucial, you can try qualcomm dm and uploading a non signed file into the efs
19:02 phh slight: put the stylus in to go to portrait mode, then long press on home (aka menu) key you'll get a virtual keyboard
19:02 slight can't work out how to enter the character as android doesn't support the 'symb' thing from winmo
19:02 tmzt shitf-. ?
19:02 tmzt or use better term
19:02 slight ah great, thanks
19:03 tmzt does it work?
19:03 tmzt no, the keymap doesn't support fn yet I think
19:03 tmzt maybe update your .kcm
19:03 slight heh can't get the bugger into portrait now :p
19:03 slight sec
19:05 slight can't get the virtual kb but my capacitive pad has been a bit dodgy in general so i'll give it a try later. thanks
19:05 slight btw phh, the build you posted today seems quite a lot more responsive
19:06 phh I posted a build ... ?
19:06 slight on xda that is
19:06 slight oh
19:06 tmzt what device?
19:06 slight no sorry :p
19:06 phh slight: ah.
19:06 slight it's babijoee posting them, brainfart
19:06 slight raph100
19:07 tmzt it's resisitve
19:07 phh cr2: ARM9 didn't suicide \o/
19:07 phh tmzt: capacitive *pad* :p
19:07 slight interesting
19:07 tmzt \o_/
19:07 slight ah yes
19:07 tmzt ah
19:07 tmzt for the menu key?
19:07 tmzt hold menu
19:07 phh cr2: ok sorry you'll still wait a bit, I try to divert the other RPC's to "correct" versions :p
19:08 tmzt \o./
19:08 tmzt phh: binary patching?
19:08 phh tmzt: naahh, in kernel, rpc command patching :p
19:08 cr2 tmzt: if it works without xtra, then xtra is nt critical
19:08 phh argh I didn't do my dump correctly
19:08 slight yeah i held menu but no dice. cleaned it and so on
19:09 tmzt what kernel on what hardware?
19:09 phh htc-msm-2.6.27 on diam100
19:09 tmzt cr2: the ttff is almost zero with xtra
19:09 tmzt on my tp2
19:09 phh what's the second f ?
19:09 tmzt it's five minutes outside without it
19:09 tmzt time to first fix
19:09 tmzt cold start vs. warm start
19:10 cr2 tmzt: we have a crash problem to solve first
19:10 tmzt it's about a second outside in google, 30 inside near window
19:10 phh cr2: I have the two first command sent by android, want it ?
19:10 tmzt I think nmea output is wrong because of no fix
19:10 tmzt but ok
19:10 tmzt wrong/null
19:11 phh cr2: sorry for the endian.
19:11 phh (again the version is fixed by the kernel to 0)
19:13 cr2 51c92bd8
19:13 cr2 is the ATL version
19:13 cr2 the VERS is all over the .so
19:13 phh whose atl version ?
19:14 phh for us it's still 0 I guess ?
19:14 cr2 #define RPC_PDSM_ATL_PROG 0x3000001d
19:14 cr2 yes, for 5225 =0
19:14 phh ok
19:15 cr2 f310fffff constant
19:16 cr2 this one is on 5225
19:17 slight thanks for the help o/
19:22 cr2 phh: looks strange
19:22 phh what ?
19:22 cr2 your pastebin
19:22 phh why ?
19:23 cr2 what did you print out ?
19:23 phh hum arm9 crashed, I guess ATL commands has to be changed too.
19:23 phh cr2: raw commands sent by libgps
19:23 cr2 yes, at least the version
19:23 phh the version is already done
19:23 cr2 hmm
19:24 phh cr2: but it's printed in little endian while it's sent as big endian
19:24 phh wait. where did they put the version id ?
19:24 tmzt (I think i2c is dying on tp2, optical is going crazy)
19:25 tmzt (light sensor)
19:25 phh tmzt: you can read it with debugfs/micropklt_dbg/light
19:25 tmzt hmm I will, I'm in ce though
19:25 tmzt on tp this was part of keyboard dying and how I got tp2
19:26 phh sounds like microp's dying
19:26 phh hum no they are two different microp
19:26 tmzt right, keyboard is fine
19:26 tmzt thankfully
19:36 phh wow
19:36 phh it did a lot of RPCs before crashing
19:38 phh cr2:
19:39 cr2 2 with payload 1
19:40 cr2 3, 1
19:40 phh and it crashes on 1d
19:40 cr2 phh: is it possible to dereference the rpc ept name ?
19:41 phh cr2: you mean translate 3000005b ?
19:41 cr2 2, b
19:41 tmzt like nslookup through rpcr?
19:41 cr2 dst=1:c3a88320
19:41 cr2 ept->name
19:41 cr2 for c3a88320
19:42 cr2 3,b
19:42 cr2 i'll go out and buy some beer. bbl
19:43 phh arch/arm/mach-msm/smd_rpcrouter.c:663: error: 'struct msm_rpc_endpoint' has no member named 'name'
19:43 phh hum.
19:43 phh there is no such thing like name in this struct
19:44 phh so the problem is that it calls a 1d proc, which doesn't exist. bah.
19:44 phh diner, bbl too.
19:53 tmzt 1d?
19:53 tmzt that's the PROC address?
19:59 cr2 phh: you trap and rewrite the calls ?
20:00 cr2 1d,3 -> 1d,4 ?
20:00 cr2 is it right ??
20:01 IceBone babijoee in here?
20:03 cr2 a9710411 these are hardcoded into
20:06 cr2 pdsm_atl_l2_proxy_reg (1,4,5)
20:06 cr2 you call pdsm_atl_l2_proxy_reg (1,a9710411,a97103f1)
20:08 cr2 5b, 4 (50a, 4ceda0, a970bf81, 00000000, f310ffff, a970bee1)
20:08 cr2 0x5,6 pdsm_client_pd_reg (clientID=da3,0,0,0,f310ffff,-1)
20:10 cr2 ok, i'll edit and pastebin it
20:10 tmzt xxd -r ?
20:14 cr2 he can print it properly
20:14 cr2 but since he is not here, i'll edit manually
20:15 tmzt oh I thought you were binary patching the .so
20:15 tmzt this is all about porting a closed driver from android to ce?
20:15 tmzt ce device
20:15 cr2 the first number is xid=, the next 2 are not interesting, then 4x0 can be dropped too
20:15 cr2 tmzt: to linux gps driver
20:16 tmzt for what device?
20:16 tmzt oh, so in c?
20:16 cr2 yes
20:16 tmzt okay cool
20:16 cr2 can be used ob g1 too
20:16 tmzt with different vers?
20:17 cr2 depending on amss ver
20:20 cr2 ok, let's compare to 5225
20:21 cr2 pdsm_client_init (pd=2/xtra=b/lcs=4)
20:21 cr2 =1 for g1
20:21 cr2 unknown
20:21 cr2 =b is xtra, ok
20:21 tmzt hmm, xtra could be the mode xtra should be in
20:22 tmzt internal or external stack
20:22 tmzt lcs is privacy setting?
20:22 cr2 some local search. read the qcom patents
20:23 cr2 pdsm_atl_l2_proxy_reg (1,4,5)
20:23 cr2 pdsm_atl_l2_proxy_reg (1,a9710411,a97103f1) got g1
20:24 cr2 pdsm_client_pd_reg (clientID=da3,0,0,0,f310ffff,-1)
20:25 phh [20:05:10] <cr2> phh: you trap and rewrite the calls ?
20:25 cr2 pdsm_client_pd_reg (50a, 4ceda0, a970bf81,0,f310ffff,a970bee1)
20:25 cr2 phh: can you parse them better ?
20:26 phh maybe
20:26 cr2 drop 1)time
20:26 tmzt cr2: Pixi sources might be up and have a different apparoch to hsuart control from userspace
20:27 phh time is linux feature, I can't really drop it, but I can delete it with awk if it's a problem for you
20:27 cr2 phh: drop val[1] and val[2]
20:27 phh cr2: I don't "translate" pdsm_atl* since I don't know what they are supposed to do
20:27 cr2 and val[6,7,8,9]
20:27 phh uh ? 6,7,8,9 are useless ?
20:27 cr2 phh: i'll tell you what to translate now
20:28 cr2 they are 0
20:28 phh oh right.
20:28 cr2 btw, do you parse responses ?
20:29 phh hum no, I'll add it
20:29 cr2 nd don't print the sequency index
20:30 phh you mean the 00000004 in "[ 219.117370] 00000004=00000000" ?
20:31 cr2 [ 219.117370] 00000004= can go away
20:31 phh if i remove 1/2/6/7/8/9 that won't really help reading, but ok.
20:31 cr2 print it in one line
20:32 cr2 i'll paste the output
20:33 phh cr2: so, should 3000005b proc 0x1d be redirected to something else ?
20:33 phh (0x1e ? :D)
20:33 cr2 no
20:33 cr2 but you need to edit the parameters
20:34 phh ouch
20:34 cr2 to match wince
20:34 cr2 there are only 3 calls anyway
20:34 cr2 all in wiki, btw
20:34 phh ok
20:35 phh but where ?
20:35 cr2 ah
20:35 cr2 3000005b 00000000 0000001d
20:35 cr2 you mean this one ?
20:35 phh yes
20:35 cr2 3000005b 00000000 0000001d
20:35 cr2 a970bdc1 00001f40 004bc7e0 00000001 000000dc a770d292 000000b4 00000001 00000000
20:36 cr2 i'll check what rhod has there.
20:37 cr2 phh:
20:37 cr2 phh: and the PROG VERS FUNC triple may be parsed with names from wiki
20:37 phh ok
20:38 cr2 the pdsm_init(1) dies not exist on wince. wince calls pdsm_init(2) and so on.
20:39 cr2 maybe add these values to wiki too ?
20:39 dcordes-kais cr2 what about htc-linux munich 2010 international congress?
20:39 phh cr2: "these" ?
20:39 cr2 dcordes-kais: leviathan agrees
20:40 leviathan yes
20:40 tmzt at this point planning to come but it's a long way and I don't know my resources
20:40 cr2 dcordes-kais: will you be at fosdem ?
20:40 leviathan I've asked FOSDEM also to make a presentation but I fear its too late
20:40 dcordes-kais cr2 where is that?
20:40 cr2 brussel
20:40 leviathan brussel
20:41 cr2 dcordes-kais: i may pick you up
20:41 cr2 it's the weekend
20:41 phh cr2: do you want anything else parsed than function's name ?
20:41 dcordes-kais cr2 ok I'm in :)
20:41 cr2 ok.
20:42 cr2 phh: no. you may add raph values though.for comparison
20:43 phh can be done afterwards
20:43 cr2 leviathan: i have found my htc_acoustic notes.
20:43 cr2 leviathan: do you know what is CE ?
20:43 dcordes-kais leviathan you also coming? any moko ppl?
20:43 leviathan hmm
20:43 leviathan I fear its too late
20:44 leviathan I wanted to first time present my project idea
20:44 leviathan I wrote an email
20:44 leviathan but they didnt answer
20:44 dcordes-kais call?
20:45 dcordes-kais br
20:45 dcordes-kais b
20:46 cr2 phh: hm. on rhod it's done in a different way ;)
20:46 phh cr2: you mean there is no 0x1d call at all?
20:47 cr2 phh: sorry
20:48 phh cr2: can you find in what do this call ?
20:48 cr2 1d on rhod is pdsm_xtra_query_data_validity()
20:48 cr2 i'll check raph
20:49 cr2 i didn't see this function on raph, but maybe i've missed it.
20:50 phh maybe we should do these tries on rhod instead of my diam then ?
20:50 tmzt do you use that?
20:50 tmzt I think raph uses dm for xtra
20:50 phh tmzt: android does.
20:50 tmzt dmrouter
20:50 cr2 phh: rhod does not need the +1 offset in rewrite
20:51 phh cr2: I know
20:51 phh but still, such tracing would be easier on an AMSS that looks more like android's one
20:54 cr2 phh: hmm. it's also 1D on raph
20:54 phh uh ? why does that stupid ARM0 crashes /o\
20:55 phh ARM9*
20:55 cr2 arm9
20:55 cr2 phh: the params are different
20:56 cr2 phh: even the init is called =1
20:56 tehtrk_home anyone ever try changing ext2 system and data filesystems to ext3 full journal?
20:56 cr2 msm rpc seems to be very fragile
20:57 cr2 tehtrk_home: if it's not root
20:57 tmzt tehtrk_home: tune2fs -j fs
20:57 tehtrk_home I'm going to give it a shot, I think it will speed reads up
20:57 tehtrk_home yeah
20:57 tmzt no reason it shouldn't work
20:58 tehtrk_home what's the rc file with all the mounts and chmods?
20:58 tehtrk_home init.rc I think?
20:58 tmzt yes
20:58 tehtrk_home can you use filesystem opetions in that as if it were an fstab?
20:59 tehtrk_home *options
20:59 tmzt with -o I think why?
20:59 tmzt oh
20:59 tehtrk_home since I can't seem to find an fstab :)
20:59 tmzt depends on android mount command
20:59 cr2 tmzt: my raph whines that the quickgps data are not uptodate
20:59 tehtrk_home yeah, gotta put data=journal to do what I'm doing
21:00 tmzt thought you meant tune2fs
21:00 tmzt don't do that
21:00 tmzt hmm
21:00 tehtrk_home can't put kernel options in default.txt for that
21:00 tehtrk_home I can use tune2fs to create the journal
21:00 tmzt maybe but that option might not be supported
21:00 tehtrk_home like you said
21:00 tehtrk_home if not I'll compile it in :P
21:00 tmzt for fs's? try but I don't know
21:01 tmzt if it's fs option rather than module it won't work
21:01 tehtrk_home yeah it's a fs option
21:01 tehtrk_home not module
21:01 tmzt or just use proper mount
21:01 tmzt fron init.rc
21:01 tehtrk_home eh? but how else would one specify ext3 journal mode
21:02 leviathan I hope I get a working msm SHR version again soon
21:02 leviathan >_<
21:02 leviathan e-wm always fucks up, when I booted up
21:02 tmzt tehtrk_home: use a proper linux utils mount not android's
21:03 tehtrk_home so, create my own fstab reflecting init.rc
21:03 tehtrk_home and comment out init.rc related lines
21:03 tmzt no
21:04 tmzt hmm
21:04 tmzt with /bin/mount you could
21:04 tehtrk_home oh just make my own script
21:04 tmzt or just busybox mount -o opts in init.rc
21:04 tehtrk_home yeah good idea
21:04 tehtrk_home hmm
21:04 tmzt fstab with /bin/mount -a would work though
21:04 tmzt or just switch to debian and help get the phone working
21:04 tmzt umm
21:04 tehtrk_home eh?
21:04 tmzt never mind
21:05 tmzt what are you working on?
21:05 tehtrk_home debian on the phone??
21:05 tmzt g1?
21:05 tehtrk_home lol that would be cool, but I need to use it as a phone too
21:05 tehtrk_home kaiser
21:05 tmzt hmm, it should work
21:06 tmzt port asoc_qdsp
21:06 tmzt then try fso or the stuff I'm working on (merphone) when it's ready
21:06 tmzt talk to gnutoo about asoc_qdsp
21:07 cr2 phh: it's an exported function
21:15 polyrhythmic tehtrk_home: many of the the G1 guys use ext3 or ext4
21:15 polyrhythmic tehtrk_home: I've been thinking about upgrading to that also
21:22 cr2 is open source ?
21:28 phh arf I have to divert the reads too
21:28 cr2 yes
21:29 cr2 hm. why does does read_audiopara_from_fiile ??
21:29 phh cr2: why not ? it actually reads audiopara from a file :p
21:29 phh like windows does
21:30 tmzt cvs?
21:30 tmzt csv
21:30 phh tmzt: yup.
21:31 phh who says they copy-pasted the code from wince ? :D
21:33 phh cr2: the answers are of type rpc_request_hdr too ?
21:34 cr2 no, a bit different i think
21:34 cr2 dump them raw first
21:34 phh ok
21:35 herpez hi. from a device without known mtype, how can we know it? (for example htc athena)
21:36 dcordes herpez, hi. what do you mean, knowing it?
21:36 dcordes herpez, you mean creating an mtype?
21:36 phh cr2: I assume all reads should be 0 ?
21:37 phh should *not*
21:37 phh because that's what I get.
21:37 dcordes cr2, we can team up a bit with the OE persons at fosdem
21:37 phh haha, I bypassed the same bytes as write
21:37 herpez dcordes - i mean, people at athena foruns are asking what mtype should they put, and i started to wonder if there is any specific mtype or if they need to create on
21:38 phh hi babijoee
21:38 babijoee hi phh :)
21:38 babijoee whats going on
21:38 dcordes herpez, 1289 htcathena
21:40 IceBone Hey, babijoee!
21:40 IceBone Great work on Eclair!
21:40 phh babijoee: I had something to tell you. you have no clue what I guess ? :p
21:41 dcordes herpez,
21:42 herpez thanks dcordes, missed that
21:42 cr2 herpez: haret knows athena
21:42 babijoee phh: hrrmm we have flashable android :)
21:42 babijoee ?
21:42 cr2 herpez: athena is stuck at the microdrive driver.
21:42 tmzt on?
21:42 phh babijoee: no
21:43 babijoee opengles
21:43 babijoee :)
21:43 phh no
21:43 babijoee camera?
21:43 phh no
21:43 babijoee extra ram
21:43 phh maybe
21:43 GNUtoo tmzt, hi
21:43 herpez cr2 - yes, there was a guy asking me and i started to read it and i saw they started in 2008 well but they are stuck long time ago, thanks!
21:43 IceBone Blowjob app?
21:43 cr2 herpez: there is 2.6.31 kernel
21:44 phh babijoee: MrPippy did some tries, more or less succesfull, try kernels prior its revert
21:44 Boydell i saw the bank2 commits and reverts, what went wrong, i thought it was working?
21:44 phh Boydell: I don't know
21:44 cr2 herpez: i can even help with it
21:44 babijoee the glemsom autobuilds were not working so i couldnt try it
21:44 IceBone Quicker wakeup on incoming call?
21:44 cr2 herpez: because i have athena too, and some thing are still not in the wiki
21:44 MrPippy my first test with bank2, it worked great, stable for a full hour doing really intensive stuff
21:45 Boydell and then?
21:45 MrPippy after that, i just couldn't get it stable no matter what the memory config
21:45 phh strange
21:45 herpez cr2 - i will send "them" to here, they will have more support. I think that with news from raph and so on, android is waking up people (just my guess)
21:45 Boydell hm
21:45 MrPippy and the panics were fundamental memory stuff
21:45 Boydell that sucks, was looking forward to getting rid of some of these panics
21:46 phh MrPippy: and you don't know what you did wrong when it worked ? :D
21:46 MrPippy i'm pretty sure i just need to backport some more patches
21:46 MrPippy s
21:46 MrPippy in
21:47 MrPippy since 2.6.27 dates from aug2008, and this patch was in sep2009, a lot happened (including a few patches specifically for sparse)
21:47 herpez in last kernels, in blackstone, the wake time is about 5 seconds. Just to inform, don't know if it exists also in other devices and if you are aware of it.
21:48 phh grrrrr
21:48 phh why all my reads are 0
21:51 cr2 phh: you must see the 0x50a response from pdsm_client_init(1)
21:51 babijoee live wallpaper is so good :)
21:51 Boydell its nice, but I heard it EATS the battery alive
21:52 phh Boydell: bah only when you display home no ?
21:52 phh it prevents the system from sleeping too
21:52 Boydell phh: why would it prevent sleep?
21:52 phh Boydell: I mean idle
21:52 tmzt wakelocks
21:52 tmzt but yeah
21:53 phh tmzt: I hope they aren't that stupid
21:53 Boydell ah
21:53 tmzt it's only on home and it times out
21:55 Boydell I hate it when my when goes into idle anyways
21:55 Boydell If I want it to sleep, I make it sleep ;)
21:56 phh Boydell: I think you don't know what idle is
21:56 Boydell maybe I'm mistaken?
21:56 phh idle is that the cpu (and only the cpu) is mostly stopped when it does nothing
21:56 tmzt idle is anytime a prcoess isn't doing something
21:56 tmzt on x86 it's hlt
21:56 tmzt the cpu sleeps until something wakes it
21:57 tmzt either timer, or in NOHZ world the next scheduled event
21:57 Boydell Ohhhhh
21:57 Boydell I was thinking after an idle time it goes into idle
21:57 Boydell *sleep
22:19 Boydell it went dead in here?
22:20 phh no, I'm just trying to dump rpc reads, but I won't stop doing mistakes
22:21 Boydell ah, so your working hard, carry on then! lol
22:21 phh hard but bad.
22:21 Boydell hey, its more than im doing right now
22:22 phh lol
22:26 Boydell could be reading "C ALL-IN-ONE DESK REFERENCE FOR DUMMIES" like me ;)
22:26 NetRipper dcordes, you rang
22:26 NetRipper ?
22:26 NetRipper :)
22:28 dcordes NetRipper, yea I intented to ask about leo news
22:28 tehtrk_home polyrhythmic: I just wonder how much of a kernel time CPU hit the journaling will incur
22:29 NetRipper cr2, to use the graphics driver from the codeaurora tree, i'd need to use everything related too? mdp? mddi?
22:29 NetRipper dcordes, ah umm ok well i kinda try to keep my blog up to date
22:29 NetRipper
22:29 NetRipper i basically havent gotten much further.. i'm trying to fix some things that still go wrong, in order to get msm_fb to work
22:30 NetRipper so what you read there is pretty accurate still
22:30 dcordes very nice
22:31 phh cr2: I'm still not sure if my read dumps are right
22:35 cr2 NetRipper: i didn't compare codeaurora with nexus tree
22:35 cr2 phh: you run this on diam100 ?
22:35 phh yes
22:36 cr2 [ 215.042449] oncrpc read
22:36 cr2 [ 215.042449] 00000001 00000001 00000000 00000000 00000000 00000000 0000050a
22:36 NetRipper cr2, the main tree of codeaurora is pretty interesting.. there's lots of comments in the drivers.. most if it is 6 months old though
22:37 cr2 this the init handle response
22:37 cr2 [ 215.042449] oncrpc read
22:37 cr2 [ 215.042449] 00000001 00000001 00000000 00000000 00000000 00000000 0000050a
22:37 NetRipper well ok might be exagurated (or however you spell it), but look at;a=blob;f=drivers/video/msm/mdp.c;h=027990d394f915f78d1793de16bebdc6d48bb13c;hb=HEAD
22:38 cr2 50a for (1)
22:38 cr2 [ 215.538238] oncrpc read
22:38 cr2 [ 215.538269] 00000001 00000001 00000000 00000000 00000000 00000000 00001f40
22:39 cr2 1f40 for (b), it's xtra and already known.
22:39 ToAsTcfh phh: i take it u didnt make it very far on the camera
22:39 phh ToAsTcfh: anyway it won't be of any help for you
22:39 phh I'll use control0 and cie
22:39 ToAsTcfh errr
22:40 phh cr2: isn't there supposed to be a prog id in the response ?
22:40 ToAsTcfh i need the libcamera to bwe writen for our camera. ive been looking at the source for it and playin a bit but its really of no use for my camera
22:40 cr2 phh: some calls get 1 and others 0 as response.
22:41 cr2 phh: yes, and it works
22:41 cr2 wince never calls =1
22:41 cr2 only =1
22:41 cr2 only =2
22:41 phh ah.
22:41 cr2 and the =b value is ok.
22:41 ToAsTcfh i cant load msm_camera because the symbols arent in the kernel
22:41 cr2 i forgot if 1 or 0 is success
22:41 cr2 check the source
22:42 tmzt what symbols?
22:42 ToAsTcfh is there another way to load modules without being able to modify the kernel
22:43 phh cr2: 0 means working
22:43 phh I think
22:43 ToAsTcfh ?
22:44 cr2 [ 216.041748] 00000002 3000001d 51c92bd8 00000003 00000001 a9710411 a97103f1 ATL callafter
22:44 cr2 [ 216.041992] 00000002 3000001d 00000000 00000004 00000001 a9710411 a97103f1
22:44 cr2 [ 216.042266] [RPC] CALL to 3000001d:0 @ 0:00000001 (52 bytes)
22:44 cr2 [ 216.040588] oncrpc read
22:44 cr2 [ 216.040618] 00000001 00000001 00000000 00000000 00000000 00000000 00000001
22:44 cr2 fails, ret=1
22:44 cr2 the last value
22:45 cr2 put 1,4,6 as params
22:45 cr2 00000001 a9710411 a97103f1 now
22:45 cr2 do
22:45 cr2 1 4 6
22:46 phh 4 instead of a971041 ?
22:46 cr2 yes
22:46 phh ok
22:46 ToAsTcfh phh: i remember u telling me i could load a module without modifying the kernel source. what were u talking about?
22:46 phh ToAsTcfh: faking the kernel identifier string
22:46 cr2 tmzt: this xtra here is PITA
22:47 ToAsTcfh hmm and thats what it called
22:47 polyrhythmic tehtrk_home: me neither, but let me know if you try it, I have a kaiser also
22:48 cr2 phh: 02:13:12 [D:GPS] GPS: [StartWorkerThreadIfNeeded] TCXO_LOCK_ID_GPS Locked
22:49 cr2 phh: i think it protects against the clock switch
22:49 tehtrk_home polyrhythmic: will do. I know you do, I'm using your eclair build from xda :P
22:50 cr2 GPS: [ProcessGpsFixRequest] operation_mode = 1
22:50 cr2 this is interesting
22:51 cr2 02:13:12 [D:GPS] GPS: [PDEvent] clientID=3491, event=32
22:51 cr2 02:13:12 [D:GPS] GPS: [PDEvent] PDSM_PD_EVENT_BEGIN
22:52 cr2 [RPC] r 51248170
22:52 cr2 02:13:12 [D:SMD] +SerOpen.
22:52 cr2 hmm
22:52 cr2 clientID=3491
22:53 cr2 what was the smd channel for xtra data ?
22:54 phh hum it's done over smd ?
22:54 polyrhythmic tehtrk_home: glad to hear it, hope to have an AOSP build out soon :) just busy with work
22:54 cr2 it's transfered somehow
22:54 cr2 phh: it can be seen in the dumps
22:54 cr2 with the xtra urls
22:55 tehtrk_home Understandable. That's one of the reasons for me not having my own build yet. AOSP?
23:00 phh cr2: oh wait. the call that crashes is procedure id=0x3 (amss 6120) or id=4 (amss 5200) but it seems they don't exist ? (or they are not documented)
23:01 phh ah it's ATL not PDAPI
23:01 phh sorry
23:01 phh cr2: am I supposed to do procedure++ for ATL ?
23:01 cr2 check wiki
23:02 phh right
23:02 cr2 4,3 pdsm_atl_l2_proxy_reg (1,4,5) g1:(1,a9710411,a97103f1)
23:02 cr2 i think you do it right already
23:04 phh hum, you told me to do 4 6
23:04 phh so 4 6 or 4 5 ?
23:05 IceBone Hey, guys... do you know what might be causing Eclair so long to wake up? It takes at least 5 rings during an incoming call to turn the screen on so I know who's calling me.
23:06 phh IceBone: mmc is slow to wake up.
23:06 IceBone mmc?
23:06 phh SD
23:06 IceBone Ah.
23:08 IceBone Can that be fixed or is it a flaw with the Diamond?
23:08 stinebd what's new? have camera, gps, gl magically started working yet?
23:08 phh stinebd: naa
23:08 tmzt that needs nand then?
23:08 phh IceBone: we will need to use nand on ebi.
23:09 stinebd gps can wait anyway, we have geolocation in the ril now
23:09 IceBone Well, today's build is much faster than that of 2 days ago.
23:09 tmzt lac?
23:09 phh stinebd: I won't driver with tower location :p
23:09 IceBone Great job, everyone.
23:09 phh tmzt: yes
23:09 stinebd phh: left turn in 4km :D
23:09 tmzt cool, that's enough for google maps?
23:09 phh tmzt: yes
23:09 phh tmzt: but seriously not for navigation.
23:10 stinebd works nice for local search results and stuff
23:10 ToAsTcfh phh: this seems to be a pretty deep quest. faking kernel identifier string isnt i guess done by meer mortals lol
23:10 stinebd and the weather and news applet if you have legally obtained it somehow
23:10 IceBone tries maps
23:11 tmzt ToAsTcfh: huh?
23:11 ToAsTcfh in other words i cant find anything on doing it.
23:11 cr2 phh: 4 5
23:11 ToAsTcfh on how to do it i meen
23:11 phh cr2: baaaaah
23:12 tmzt what string?
23:12 tmzt uname's?
23:12 stinebd phh: you haven't by chance tried out nohz yet have you?
23:12 phh stinebd: no
23:12 tmzt look for sys_uname I think
23:12 tmzt but I wouldn't fake that
23:12 cr2 stinebd: with opencelldb ?
23:12 stinebd phh: i'm having trouble finding faults in it, aside from losing adb shortly after boot and idle sleep
23:12 ToAsTcfh when loading a module modprobe needs things from the kernel its self.
23:13 stinebd cr2: was that meant for me?
23:13 cr2 stinebd:
23:13 stinebd cr2: i don't understand
23:13 phh cr2: no, google's DB
23:13 ToAsTcfh the kernel i guess needs to have symbols added to know what to do with the new found module
23:14 stinebd ohhh for geolocation
23:14 stinebd cr2: phh is right, google's
23:14 chamonix cr2 and stinebd : besides opencellid you may want to try ericsson's cell-id lookup too:
23:14 phh chamonix: uh ? ericson's offers his DB ?
23:14 chamonix yea
23:14 chamonix 3,6M cells
23:15 stinebd might i ask why to do this?
23:15 cr2 chamonix: i want to gather such info in linux
23:15 phh stinebd: google ?
23:15 chamonix cr2: how do you mean "gather"? you want to log?
23:15 phh chamonix: explore the entire world with GPS+LAC
23:15 phh and do a DB with that
23:15 phh a free db
23:16 stinebd that could take some time
23:16 phh like OSM
23:16 chamonix well there are quite a few of those already, with JSon and webservice interface
23:16 chamonix googlemaps has a reverse engineered API too
23:16 chamonix I men online
23:16 chamonix s/men/mean
23:16 phh chamonix: an open DB i meana.
23:16 phh -a
23:16 cr2 chamonix: yes. when i'm close to a tower, i'dlike to know its data
23:17 chamonix cr2: I can send you my code that does that
23:17 chamonix using telefonymanager + whateverwebservice or gmaps
23:17 cr2 chamonix: code ?
23:17 chamonix java code
23:18 cr2 you are missing the point
23:18 chamonix I guess :)
23:18 cr2 i know my location with gps
23:18 phh I'm definitely a bad coder
23:18 chamonix ok so u want surrounding cells right?
23:19 cr2 so i'd like to store the base station data like LAC/CID etc
23:19 cr2 for gsm we can do itwith at+creg=2 and at+csq/at+htccsq
23:19 cr2 what about umts
23:19 cr2 ?
23:19 phh cr2: same
23:19 phh I don't know why, but it works the same way
23:20 chamonix me neither
23:20 phh (afaik it's not supposed to)
23:20 chamonix it's not supposed to :)
23:20 cr2 there are many debug AT commands, but they are undocumented
23:20 cr2 phh: i get CID=0
23:20 chamonix UMTS has psc for surrounding cells
23:20 phh cr2: really ?!?
23:20 cr2 and fixed CSQ
23:20 tmzt phh: openbmap has that mostly
23:20 cr2 yes
23:20 tmzt isn't lac related to wgs somehow?
23:20 phh wgs ?
23:21 tmzt ToAsTcfh: what symbols, pleasr list them
23:21 tmzt a way of encoding location on the planet
23:22 chamonix lac it's not geolocation
23:22 phh afaik it's not related at all
23:22 cr2 phh: in your log only this atl call failed
23:22 cr2 tmzt: geolocation=lat/long/z
23:22 cr2 phh: the first 1,4,5 call
23:22 chamonix but u can get geolocation from lac + fancy algorithm + database
23:22 phh chamonix: uh ? you need algos ?
23:23 phh luc_: no it's not
23:23 cr2 chamonix: where do i download this database for cote d'ivoir ?
23:24 chamonix phh: as far as I got it there is some triangulation when multiple cells are available
23:25 chamonix cr2: wow, I haven't seen any downloadable data, just webservices
23:25 phh chamonix: just tell me how to get multiple cells at the same time :D
23:25 cr2 chamonix: it's the same idea as openstreetmap. such databases contain a lot ofoutdated, wronganf intentionally false information
23:25 chamonix phh: RIL lists the neighbouring cells
23:25 phh chamonix: no it doesn't
23:25 chamonix well, 1 at least :)
23:25 cr2 chamonix: with wihc call ?
23:26 phh chamonix: I checked with htc ril, it does the same
23:26 tmzt cr2: when are you going there?
23:26 cr2 phh: at+3gncell ?
23:26 phh cr2: I've never seen it returning anything
23:26 chamonix requestNeighboringCellIds
23:26 cr2 tmzt: was just n example
23:26 phh chamonix: yes that's theorically :p
23:26 cr2 phh: not on 5225 amss:(
23:27 phh cr2: ok
23:27 cr2 phh: but there are at$3g* commands
23:27 phh yes
23:27 phh I tried them without much luck
23:27 cr2 phh: some of them work
23:28 cr2 phh: and there is a filedtest.exe
23:28 chamonix I've read the UMTS spec the other day, there's noting much except tons of theory and timings
23:28 cr2 phh: you can dump AT smd fifo while using it
23:28 phh cr2: iirc I did
23:29 cr2 chamonix: some guy fron TX posted some helpfl doc at xda-dev
23:29 chamonix regarding RIL or location stuff?
23:29 phh lol, dumping ramconsole crashes wince
23:30 ToAsTcfh tmzt: hold up i think i got it. modprobe passes a string intern it gets an identifier. then modprob checks the modprobe.conf for that string. then knows to load the .ko. i think this is what phh was talking about. if im wrong phh please let me know
23:30 phh ToAsTcfh: not at all :p
23:30 phh it's the kernel who checks the string
23:30 phh the 2.6.27 ARMEv6 blabla
23:30 ToAsTcfh ok ok my bad.
23:30 cr2 omg. geoid separation
23:30 chamonix I must admit I use loc only so my mobile knows where it is to switch to the right profile.... I hate when my phone rings while I'm in a meeting so it turns to silent whenever I'm near the office :)
23:31 phh chamonix: I need that.
23:31 RandomPPCGeek hey all
23:31 ToAsTcfh yeah but if the kernel doesnt know the string then how to monipulatte that even in my favor
23:31 cr2 must be veryprimitive though
23:32 chamonix phh: I'll publish my code when I'm finished but you can have a look at "locale beta" from the market.... it's pretty nice except it uses only the google api and you won't see any technical stuff like cid+lac
23:32 stinebd cool
23:32 stinebd thanks chamonix
23:32 phh I knew I've seen such an app on the store.
23:33 stinebd i saw an app that makes noises of various bodily functions
23:33 chamonix right now my LocService detect position, charging, time, and adjusts to any profile based on rules, and sends stuff like tweets or SMS when you enter an area e.g.
23:33 phh huhu
23:34 IceBone The only thing that needed autotweeting was the newlyweds' bed.
23:34 chamonix phh: locale won the andoir innovation contest, there's the beta and a paid version but the beta is maintained and it has scripting features plus a really cool API
23:34 chamonix s/andoir/android
23:34 stinebd so you're saying don't buy it
23:34 cr2 chamonix: sound like perfect spyware :)
23:34 chamonix lol
23:34 chamonix cr2: have a look at google lattitude
23:35 stinebd it's not spying if you opt in ;)
23:35 cr2 yah, theyshouldburninhell
23:35 chamonix stinebd: I would pay for it but I just wanted to says it's not just a beta to attract ppl and "force" them paying afterward
23:35 tmzt ToAsTcfh: alias?
23:35 tmzt yeah
23:35 tmzt but not modver string
23:35 tehtrk_home polyrhythmic: does the latest eclair build seem slower? It could either be that, my partitioning, or not using the hybrid rootfs that i was using with the mid-december eclair build I was using
23:35 tmzt but you said missing symbols
23:36 phh [ 218.192718] timeout waiting for mdp to complete 1
23:36 phh what's the hell.
23:36 cr2 chamonix: goole IS the spyware business
23:36 phh :)
23:37 cr2 i draw my own maps at openstreetmaps
23:37 chamonix well, even google discovered that there's a markting argument in at least pretending to inform ppl about the data held about them
23:37 chamonix it's politically correctness
23:37 chamonix -ly
23:37 cr2 and they are much better and accurate than what google can provide
23:38 chamonix that would be cool to replace the google api by another one
23:38 chamonix as per loc I mean
23:39 polyrhythmic tehtrk_home: not to me, and I have email & google voice running in the bkgrnd... I make good use of taskiller though
23:41 chamonix cr2: <--- someone already did that, so I'm definitely going to integrate that
23:41 phh D/libgps ( 1178): qct_gps_set_position_mode = 1, fix_frequency = 1
23:41 phh D/libgps ( 1178): qct_gps_start
23:41 phh cr2: seems better.
23:41 phh kind of.
23:41 stinebd hmm
23:41 cr2 phh: show me the full log
23:41 chamonix and it's interesting, behind the API it's ericsson :-D
23:41 cr2 hah.
23:41 stinebd it appears to be in some sort of german language
23:42 chamonix sry for that stinebd :)
23:42 stinebd that's okay, i took four years of german
23:42 stinebd some day i'll remember how to read it
23:42 chamonix lol
23:42 phh cr2: just to mention, it crashed.
23:42 IceBone Anyone know where I could get the Android logo, but in nice 3D, like it is in the Hero startup animation?
23:43 chamonix no need, click on the first link and you get here:
23:43 chamonix apii key registered
23:44 cr2 chamonix:
23:44 cr2 phh: you use mode=1, which is never used on wince
23:45 chamonix cool
23:45 chamonix finally I got to code Java again :)
23:45 chamonix s/got/get
23:46 phh GRRRRRR
23:46 phh ATL is wrong AGAIN
23:46 phh why can't i code that.
23:46 phh and how can it not return an error if it's wrong
23:51 ToAsTcfh phh: so i need to create and edit /etc/modprobe.conf to have an alias for the .ko?
23:51 phh no
23:52 ToAsTcfh ok then im reading the wrong thing
23:53 ToAsTcfh or modprobe will make the .conf oncwe it gets the identifier part from the kernel
23:53 phh and i'm writing the wrong thing.
23:53 phh no.
23:53 phh it's not linked with modprobe at all
23:53 rzk__ polyrhythmic: I've uploaded latest tinboot build for kaiser, working good with your eclair build, test if you want -
23:55 cr2 phh: PDSM_PD_EVENT_*, do we know something about it ?
23:55 phh I don't.
23:55 IceBone
23:56 ToAsTcfh errr... this is frustrating. i cant even figure out how to do this. or im just looking in all the wrong places
23:57 cr2 rzk__: do you have working gps on kaiser ?
23:57 rzk__ cr2: i have no kaiser, lol.
23:57 rzk__ I have niki
23:58 cr2 it's the same for me :)
23:59 rzk__ gps is hardly disabled on niki, no code in radio & spl, someone on forum tried to figure out what is deleted, but no news from him.